id	sid	tid	token	lemma	pos
bracis-33593	1	1	likelihood	likelihood	NOUN
bracis-33593	1	2	estimator	estimator	NOUN
bracis-33593	1	3	for	for	ADP
bracis-33593	1	4	 	 	SPACE
bracis-33593	1	5	multi	multi	PROPN
bracis-33593	1	6	model	model	NOUN
bracis-33593	1	7	-	-	PUNCT
bracis-33593	1	8	based	base	VERB
bracis-33593	1	9	reinforcement	reinforcement	NOUN
bracis-33593	1	10	learning	learning	NOUN
bracis-33593	1	11	|	|	ADV
bracis-33593	1	12	springer	springer	NOUN
bracis-33593	1	13	nature	nature	PROPN
bracis-33593	1	14	link	link	PROPN
bracis-33593	1	15	(	(	PUNCT
bracis-33593	1	16	formerly	formerly	ADV
bracis-33593	1	17	springerlink	springerlink	NOUN
bracis-33593	1	18	)	)	PUNCT
bracis-33593	1	19	skip	skip	VERB
bracis-33593	1	20	to	to	ADP
bracis-33593	1	21	main	main	ADJ
bracis-33593	1	22	content	content	NOUN
bracis-33593	1	23	advertisement	advertisement	NOUN
bracis-33593	1	24	log	log	NOUN
bracis-33593	1	25	in	in	ADP
bracis-33593	1	26	menu	menu	NOUN
bracis-33593	1	27	find	find	VERB
bracis-33593	1	28	a	a	DET
bracis-33593	1	29	journal	journal	NOUN
bracis-33593	1	30	publish	publish	VERB
bracis-33593	1	31	with	with	ADP
bracis-33593	1	32	us	we	PRON
bracis-33593	1	33	track	track	VERB
bracis-33593	1	34	your	your	PRON
bracis-33593	1	35	research	research	NOUN
bracis-33593	1	36	search	search	NOUN
bracis-33593	1	37	cart	cart	NOUN
bracis-33593	1	38	home	home	NOUN
bracis-33593	1	39	intelligent	intelligent	ADJ
bracis-33593	1	40	systems	system	NOUN
bracis-33593	1	41	conference	conference	NOUN
bracis-33593	1	42	paper	paper	NOUN
bracis-33593	1	43	likelihood	likelihood	NOUN
bracis-33593	1	44	estimator	estimator	NOUN
bracis-33593	1	45	for	for	ADP
bracis-33593	1	46	 	 	SPACE
bracis-33593	1	47	multi	multi	PROPN
bracis-33593	1	48	model	model	NOUN
bracis-33593	1	49	-	-	PUNCT
bracis-33593	1	50	based	base	VERB
bracis-33593	1	51	reinforcement	reinforcement	NOUN
bracis-33593	1	52	learning	learning	NOUN
bracis-33593	1	53	conference	conference	NOUN
bracis-33593	1	54	paper	paper	NOUN
bracis-33593	1	55	first	first	ADV
bracis-33593	1	56	online	online	ADV
bracis-33593	1	57	:	:	PUNCT
bracis-33593	1	58	30	30	NUM
bracis-33593	1	59	january	january	NOUN
bracis-33593	1	60	2025	2025	NUM
bracis-33593	1	61	pp	pp	CCONJ
bracis-33593	1	62	184–198	184–198	NUM
bracis-33593	1	63	cite	cite	VERB
bracis-33593	1	64	this	this	DET
bracis-33593	1	65	conference	conference	NOUN
bracis-33593	1	66	paper	paper	NOUN
bracis-33593	1	67	access	access	NOUN
bracis-33593	1	68	provided	provide	VERB
bracis-33593	1	69	by	by	ADP
bracis-33593	1	70	university	university	PROPN
bracis-33593	1	71	of	of	ADP
bracis-33593	1	72	notre	notre	PROPN
bracis-33593	1	73	dame	dame	PROPN
bracis-33593	1	74	hesburgh	hesburgh	PROPN
bracis-33593	1	75	library	library	PROPN
bracis-33593	1	76	download	download	PROPN
bracis-33593	1	77	book	book	NOUN
bracis-33593	1	78	pdf	pdf	PROPN
bracis-33593	1	79	download	download	NOUN
bracis-33593	1	80	book	book	NOUN
bracis-33593	1	81	epub	epub	PROPN
bracis-33593	1	82	intelligent	intelligent	ADJ
bracis-33593	1	83	systems	system	NOUN
bracis-33593	1	84	(	(	PUNCT
bracis-33593	1	85	bracis	bracis	NOUN
bracis-33593	1	86	2024	2024	NUM
bracis-33593	1	87	)	)	PUNCT
bracis-33593	1	88	likelihood	likelihood	NOUN
bracis-33593	1	89	estimator	estimator	NOUN
bracis-33593	1	90	for	for	ADP
bracis-33593	1	91	 	 	SPACE
bracis-33593	1	92	multi	multi	PROPN
bracis-33593	1	93	model	model	NOUN
bracis-33593	1	94	-	-	PUNCT
bracis-33593	1	95	based	base	VERB
bracis-33593	1	96	reinforcement	reinforcement	NOUN
bracis-33593	1	97	learning	learning	NOUN
bracis-33593	1	98	download	download	NOUN
bracis-33593	1	99	book	book	NOUN
bracis-33593	1	100	pdf	pdf	PROPN
bracis-33593	1	101	download	download	NOUN
bracis-33593	1	102	book	book	PROPN
bracis-33593	1	103	epub	epub	PROPN
bracis-33593	1	104	guilherme	guilherme	PROPN
bracis-33593	1	105	albarrans9	albarrans9	PROPN
bracis-33593	1	106	&	&	CCONJ
bracis-33593	1	107	valdinei	valdinei	PROPN
bracis-33593	1	108	freire9	freire9	PROPN
bracis-33593	1	109	  	  	SPACE
bracis-33593	1	110	part	part	NOUN
bracis-33593	1	111	of	of	ADP
bracis-33593	1	112	the	the	DET
bracis-33593	1	113	book	book	NOUN
bracis-33593	1	114	series	series	NOUN
bracis-33593	1	115	:	:	PUNCT
bracis-33593	1	116	lecture	lecture	NOUN
bracis-33593	1	117	notes	note	NOUN
bracis-33593	1	118	in	in	ADP
bracis-33593	1	119	computer	computer	NOUN
bracis-33593	1	120	science	science	NOUN
bracis-33593	1	121	(	(	PUNCT
bracis-33593	1	122	(	(	PUNCT
bracis-33593	1	123	lnai	lnai	ADJ
bracis-33593	1	124	,	,	PUNCT
bracis-33593	1	125	volume	volume	NOUN
bracis-33593	1	126	15413	15413	NUM
bracis-33593	1	127	)	)	PUNCT
bracis-33593	1	128	)	)	PUNCT
bracis-33593	1	129	included	include	VERB
bracis-33593	1	130	in	in	ADP
bracis-33593	1	131	the	the	DET
bracis-33593	1	132	following	follow	VERB
bracis-33593	1	133	conference	conference	NOUN
bracis-33593	1	134	series	series	NOUN
bracis-33593	1	135	:	:	PUNCT
bracis-33593	1	136	brazilian	brazilian	ADJ
bracis-33593	1	137	conference	conference	NOUN
bracis-33593	1	138	on	on	ADP
bracis-33593	1	139	intelligent	intelligent	ADJ
bracis-33593	1	140	systems	system	NOUN
bracis-33593	1	141	361	361	NUM
bracis-33593	1	142	accesses	access	NOUN
bracis-33593	1	143	abstract	abstract	ADV
bracis-33593	1	144	efficiently	efficiently	ADV
bracis-33593	1	145	learning	learn	VERB
bracis-33593	1	146	and	and	CCONJ
bracis-33593	1	147	generalizing	generalize	VERB
bracis-33593	1	148	from	from	ADP
bracis-33593	1	149	interactions	interaction	NOUN
bracis-33593	1	150	with	with	ADP
bracis-33593	1	151	the	the	DET
bracis-33593	1	152	environment	environment	NOUN
bracis-33593	1	153	is	be	AUX
bracis-33593	1	154	a	a	DET
bracis-33593	1	155	concern	concern	NOUN
bracis-33593	1	156	in	in	ADP
bracis-33593	1	157	reinforcement	reinforcement	NOUN
bracis-33593	1	158	learning	learning	NOUN
bracis-33593	1	159	.	.	PUNCT
bracis-33593	2	1	model	model	NOUN
bracis-33593	2	2	-	-	PUNCT
bracis-33593	2	3	based	base	VERB
bracis-33593	2	4	approaches	approach	NOUN
bracis-33593	2	5	offer	offer	VERB
bracis-33593	2	6	promising	promise	VERB
bracis-33593	2	7	potential	potential	NOUN
bracis-33593	2	8	to	to	PART
bracis-33593	2	9	address	address	VERB
bracis-33593	2	10	these	these	DET
bracis-33593	2	11	challenges	challenge	NOUN
bracis-33593	2	12	,	,	PUNCT
bracis-33593	2	13	leveraging	leverage	VERB
bracis-33593	2	14	the	the	DET
bracis-33593	2	15	underlying	underlie	VERB
bracis-33593	2	16	structure	structure	NOUN
bracis-33593	2	17	of	of	ADP
bracis-33593	2	18	the	the	DET
bracis-33593	2	19	world	world	NOUN
bracis-33593	2	20	to	to	PART
bracis-33593	2	21	speed	speed	VERB
bracis-33593	2	22	learning	learning	NOUN
bracis-33593	2	23	and	and	CCONJ
bracis-33593	2	24	improve	improve	VERB
bracis-33593	2	25	generalization	generalization	NOUN
bracis-33593	2	26	.	.	PUNCT
bracis-33593	3	1	by	by	ADP
bracis-33593	3	2	assuming	assume	VERB
bracis-33593	3	3	a	a	DET
bracis-33593	3	4	structured	structured	ADJ
bracis-33593	3	5	environment	environment	NOUN
bracis-33593	3	6	,	,	PUNCT
bracis-33593	3	7	such	such	ADJ
bracis-33593	3	8	approaches	approach	NOUN
bracis-33593	3	9	aim	aim	VERB
bracis-33593	3	10	to	to	PART
bracis-33593	3	11	capture	capture	VERB
bracis-33593	3	12	key	key	ADJ
bracis-33593	3	13	patterns	pattern	NOUN
bracis-33593	3	14	from	from	ADP
bracis-33593	3	15	limited	limited	ADJ
bracis-33593	3	16	interactions	interaction	NOUN
bracis-33593	3	17	and	and	CCONJ
bracis-33593	3	18	synthesize	synthesize	VERB
bracis-33593	3	19	experiences	experience	NOUN
bracis-33593	3	20	for	for	ADP
bracis-33593	3	21	enhanced	enhanced	ADJ
bracis-33593	3	22	exploitation	exploitation	NOUN
bracis-33593	3	23	and	and	CCONJ
bracis-33593	3	24	generalization	generalization	NOUN
bracis-33593	3	25	to	to	ADP
bracis-33593	3	26	unexplored	unexplored	ADJ
bracis-33593	3	27	territories	territory	NOUN
bracis-33593	3	28	.	.	PUNCT
bracis-33593	4	1	in	in	ADP
bracis-33593	4	2	this	this	DET
bracis-33593	4	3	paper	paper	NOUN
bracis-33593	4	4	,	,	PUNCT
bracis-33593	4	5	we	we	PRON
bracis-33593	4	6	test	test	VERB
bracis-33593	4	7	a	a	DET
bracis-33593	4	8	model	model	NOUN
bracis-33593	4	9	-	-	PUNCT
bracis-33593	4	10	based	base	VERB
bracis-33593	4	11	reinforcement	reinforcement	NOUN
bracis-33593	4	12	learning	learning	NOUN
bracis-33593	4	13	framework	framework	NOUN
bracis-33593	4	14	aimed	aim	VERB
bracis-33593	4	15	at	at	ADP
bracis-33593	4	16	further	far	ADV
bracis-33593	4	17	advancing	advance	VERB
bracis-33593	4	18	the	the	DET
bracis-33593	4	19	frontiers	frontier	NOUN
bracis-33593	4	20	of	of	ADP
bracis-33593	4	21	efficiency	efficiency	NOUN
bracis-33593	4	22	and	and	CCONJ
bracis-33593	4	23	generalization	generalization	NOUN
bracis-33593	4	24	.	.	PUNCT
bracis-33593	5	1	motivated	motivate	VERB
bracis-33593	5	2	by	by	ADP
bracis-33593	5	3	the	the	DET
bracis-33593	5	4	insight	insight	NOUN
bracis-33593	5	5	that	that	SCONJ
bracis-33593	5	6	environments	environment	NOUN
bracis-33593	5	7	often	often	ADV
bracis-33593	5	8	exhibit	exhibit	VERB
bracis-33593	5	9	distinct	distinct	ADJ
bracis-33593	5	10	regions	region	NOUN
bracis-33593	5	11	with	with	ADP
bracis-33593	5	12	varying	vary	VERB
bracis-33593	5	13	dynamics	dynamic	NOUN
bracis-33593	5	14	,	,	PUNCT
bracis-33593	5	15	we	we	PRON
bracis-33593	5	16	introduce	introduce	VERB
bracis-33593	5	17	additional	additional	ADJ
bracis-33593	5	18	assumptions	assumption	NOUN
bracis-33593	5	19	about	about	ADP
bracis-33593	5	20	the	the	DET
bracis-33593	5	21	structure	structure	NOUN
bracis-33593	5	22	of	of	ADP
bracis-33593	5	23	the	the	DET
bracis-33593	5	24	world	world	NOUN
bracis-33593	5	25	to	to	PART
bracis-33593	5	26	facilitate	facilitate	VERB
bracis-33593	5	27	faster	fast	ADJ
bracis-33593	5	28	generalization	generalization	NOUN
bracis-33593	5	29	.	.	PUNCT
bracis-33593	6	1	specifically	specifically	ADV
bracis-33593	6	2	,	,	PUNCT
bracis-33593	6	3	we	we	PRON
bracis-33593	6	4	construct	construct	VERB
bracis-33593	6	5	a	a	DET
bracis-33593	6	6	model	model	NOUN
bracis-33593	6	7	comprising	comprise	VERB
bracis-33593	6	8	a	a	DET
bracis-33593	6	9	classifier	classifier	NOUN
bracis-33593	6	10	that	that	SCONJ
bracis-33593	6	11	,	,	PUNCT
bracis-33593	6	12	given	give	VERB
bracis-33593	6	13	a	a	DET
bracis-33593	6	14	state	state	NOUN
bracis-33593	6	15	input	input	NOUN
bracis-33593	6	16	,	,	PUNCT
bracis-33593	6	17	selects	select	VERB
bracis-33593	6	18	parameters	parameter	NOUN
bracis-33593	6	19	characterizing	characterize	VERB
bracis-33593	6	20	the	the	DET
bracis-33593	6	21	operating	operate	VERB
bracis-33593	6	22	regime	regime	NOUN
bracis-33593	6	23	of	of	ADP
bracis-33593	6	24	the	the	DET
bracis-33593	6	25	environment	environment	NOUN
bracis-33593	6	26	.	.	PUNCT
bracis-33593	7	1	both	both	CCONJ
bracis-33593	7	2	the	the	DET
bracis-33593	7	3	model	model	NOUN
bracis-33593	7	4	and	and	CCONJ
bracis-33593	7	5	the	the	DET
bracis-33593	7	6	regional	regional	ADJ
bracis-33593	7	7	parameters	parameter	NOUN
bracis-33593	7	8	are	be	AUX
bracis-33593	7	9	learned	learn	VERB
bracis-33593	7	10	from	from	ADP
bracis-33593	7	11	past	past	ADJ
bracis-33593	7	12	experiences	experience	NOUN
bracis-33593	7	13	using	use	VERB
bracis-33593	7	14	a	a	DET
bracis-33593	7	15	likelihood	likelihood	NOUN
bracis-33593	7	16	function	function	NOUN
bracis-33593	7	17	.	.	PUNCT
bracis-33593	8	1	leveraging	leverage	VERB
bracis-33593	8	2	this	this	DET
bracis-33593	8	3	model	model	NOUN
bracis-33593	8	4	,	,	PUNCT
bracis-33593	8	5	we	we	PRON
bracis-33593	8	6	employ	employ	VERB
bracis-33593	8	7	a	a	DET
bracis-33593	8	8	rapid	rapid	ADJ
bracis-33593	8	9	exploring	explore	VERB
bracis-33593	8	10	random	random	ADJ
bracis-33593	8	11	tree	tree	NOUN
bracis-33593	8	12	based	base	VERB
bracis-33593	8	13	planner	planner	NOUN
bracis-33593	8	14	to	to	PART
bracis-33593	8	15	generate	generate	VERB
bracis-33593	8	16	new	new	ADJ
bracis-33593	8	17	real	real	ADJ
bracis-33593	8	18	experiences	experience	NOUN
bracis-33593	8	19	,	,	PUNCT
bracis-33593	8	20	capitalizing	capitalize	VERB
bracis-33593	8	21	on	on	ADP
bracis-33593	8	22	the	the	DET
bracis-33593	8	23	identified	identify	VERB
bracis-33593	8	24	structural	structural	ADJ
bracis-33593	8	25	nuances	nuance	NOUN
bracis-33593	8	26	within	within	ADP
bracis-33593	8	27	the	the	DET
bracis-33593	8	28	environment	environment	NOUN
bracis-33593	8	29	.	.	PUNCT
bracis-33593	9	1	to	to	PART
bracis-33593	9	2	evaluate	evaluate	VERB
bracis-33593	9	3	the	the	DET
bracis-33593	9	4	efficacy	efficacy	NOUN
bracis-33593	9	5	learning	learn	VERB
bracis-33593	9	6	a	a	DET
bracis-33593	9	7	segmented	segment	VERB
bracis-33593	9	8	model	model	NOUN
bracis-33593	9	9	,	,	PUNCT
bracis-33593	9	10	we	we	PRON
bracis-33593	9	11	conduct	conduct	VERB
bracis-33593	9	12	a	a	DET
bracis-33593	9	13	comparative	comparative	ADJ
bracis-33593	9	14	analysis	analysis	NOUN
bracis-33593	9	15	against	against	ADP
bracis-33593	9	16	a	a	DET
bracis-33593	9	17	traditional	traditional	ADJ
bracis-33593	9	18	method	method	NOUN
bracis-33593	9	19	that	that	PRON
bracis-33593	9	20	employs	employ	VERB
bracis-33593	9	21	a	a	DET
bracis-33593	9	22	single	single	ADJ
bracis-33593	9	23	general	general	ADJ
bracis-33593	9	24	model	model	NOUN
bracis-33593	9	25	to	to	PART
bracis-33593	9	26	learn	learn	VERB
bracis-33593	9	27	the	the	DET
bracis-33593	9	28	dynamics	dynamic	NOUN
bracis-33593	9	29	of	of	ADP
bracis-33593	9	30	the	the	DET
bracis-33593	9	31	entire	entire	ADJ
bracis-33593	9	32	environment	environment	NOUN
bracis-33593	9	33	.	.	PUNCT
bracis-33593	10	1	our	our	PRON
bracis-33593	10	2	results	result	NOUN
bracis-33593	10	3	demonstrate	demonstrate	VERB
bracis-33593	10	4	the	the	DET
bracis-33593	10	5	superiority	superiority	NOUN
bracis-33593	10	6	of	of	ADP
bracis-33593	10	7	the	the	DET
bracis-33593	10	8	segmented	segment	VERB
bracis-33593	10	9	approach	approach	NOUN
bracis-33593	10	10	in	in	ADP
bracis-33593	10	11	terms	term	NOUN
bracis-33593	10	12	of	of	ADP
bracis-33593	10	13	both	both	DET
bracis-33593	10	14	efficiency	efficiency	NOUN
bracis-33593	10	15	and	and	CCONJ
bracis-33593	10	16	generalization	generalization	NOUN
bracis-33593	10	17	,	,	PUNCT
bracis-33593	10	18	underscoring	underscore	VERB
bracis-33593	10	19	the	the	DET
bracis-33593	10	20	benefits	benefit	NOUN
bracis-33593	10	21	of	of	ADP
bracis-33593	10	22	incorporating	incorporate	VERB
bracis-33593	10	23	additional	additional	ADJ
bracis-33593	10	24	assumptions	assumption	NOUN
bracis-33593	10	25	about	about	ADP
bracis-33593	10	26	the	the	DET
bracis-33593	10	27	structure	structure	NOUN
bracis-33593	10	28	of	of	ADP
bracis-33593	10	29	the	the	DET
bracis-33593	10	30	world	world	NOUN
bracis-33593	10	31	into	into	ADP
bracis-33593	10	32	model	model	NOUN
bracis-33593	10	33	-	-	PUNCT
bracis-33593	10	34	based	base	VERB
bracis-33593	10	35	reinforcement	reinforcement	NOUN
bracis-33593	10	36	learning	learning	NOUN
bracis-33593	10	37	paradigms	paradigm	VERB
bracis-33593	10	38	.	.	PUNCT
bracis-33593	11	1	access	access	NOUN
bracis-33593	11	2	provided	provide	VERB
bracis-33593	11	3	by	by	ADP
bracis-33593	11	4	university	university	PROPN
bracis-33593	11	5	of	of	ADP
bracis-33593	11	6	notre	notre	PROPN
bracis-33593	11	7	dame	dame	PROPN
bracis-33593	11	8	hesburgh	hesburgh	PROPN
bracis-33593	11	9	library	library	PROPN
bracis-33593	11	10	.	.	PUNCT
bracis-33593	12	1	download	download	PROPN
bracis-33593	12	2	conference	conference	NOUN
bracis-33593	12	3	paper	paper	NOUN
bracis-33593	12	4	pdf	pdf	NOUN
bracis-33593	12	5	similar	similar	ADJ
bracis-33593	12	6	content	content	NOUN
bracis-33593	12	7	being	be	AUX
bracis-33593	12	8	viewed	view	VERB
bracis-33593	12	9	by	by	ADP
bracis-33593	12	10	others	other	NOUN
bracis-33593	12	11	exploring	explore	VERB
bracis-33593	12	12	ensemble	ensemble	ADJ
bracis-33593	12	13	error	error	NOUN
bracis-33593	12	14	exploration	exploration	NOUN
bracis-33593	12	15	for	for	ADP
bracis-33593	12	16	 	 	SPACE
bracis-33593	12	17	unsupervised	unsupervised	ADJ
bracis-33593	12	18	reinforcement	reinforcement	NOUN
bracis-33593	12	19	learning	learning	NOUN
bracis-33593	12	20	chapter	chapter	NOUN
bracis-33593	12	21	©	©	PROPN
bracis-33593	12	22	2025	2025	NUM
bracis-33593	12	23	a	a	DET
bracis-33593	12	24	survey	survey	NOUN
bracis-33593	12	25	on	on	ADP
bracis-33593	12	26	model	model	NOUN
bracis-33593	12	27	-	-	PUNCT
bracis-33593	12	28	based	base	VERB
bracis-33593	12	29	reinforcement	reinforcement	NOUN
bracis-33593	12	30	learning	learning	NOUN
bracis-33593	12	31	article	article	NOUN
bracis-33593	12	32	23	23	NUM
bracis-33593	12	33	january	january	NOUN
bracis-33593	12	34	2024	2024	NUM
bracis-33593	12	35	learning	learn	VERB
bracis-33593	12	36	probabilistic	probabilistic	ADJ
bracis-33593	12	37	models	model	NOUN
bracis-33593	12	38	for	for	ADP
bracis-33593	12	39	model	model	NOUN
bracis-33593	12	40	checking	checking	NOUN
bracis-33593	12	41	:	:	PUNCT
bracis-33593	12	42	an	an	DET
bracis-33593	12	43	evolutionary	evolutionary	ADJ
bracis-33593	12	44	approach	approach	NOUN
bracis-33593	12	45	and	and	CCONJ
bracis-33593	12	46	an	an	DET
bracis-33593	12	47	empirical	empirical	ADJ
bracis-33593	12	48	study	study	NOUN
bracis-33593	12	49	article	article	NOUN
bracis-33593	12	50	25	25	NUM
bracis-33593	12	51	april	april	PROPN
bracis-33593	12	52	2018	2018	NUM
bracis-33593	12	53	1	1	NUM
bracis-33593	12	54	introduction	introduction	NOUN
bracis-33593	12	55	model	model	NOUN
bracis-33593	12	56	-	-	PUNCT
bracis-33593	12	57	based	base	VERB
bracis-33593	12	58	reinforcement	reinforcement	NOUN
bracis-33593	12	59	learning	learning	NOUN
bracis-33593	12	60	(	(	PUNCT
bracis-33593	12	61	rl	rl	NOUN
bracis-33593	12	62	)	)	PUNCT
bracis-33593	12	63	has	have	VERB
bracis-33593	12	64	a	a	DET
bracis-33593	12	65	focus	focus	NOUN
bracis-33593	12	66	on	on	ADP
bracis-33593	12	67	efficiency	efficiency	NOUN
bracis-33593	12	68	and	and	CCONJ
bracis-33593	12	69	generalization	generalization	NOUN
bracis-33593	12	70	through	through	ADP
bracis-33593	12	71	model	model	NOUN
bracis-33593	12	72	construction	construction	NOUN
bracis-33593	12	73	.	.	PUNCT
bracis-33593	13	1	by	by	ADP
bracis-33593	13	2	leveraging	leverage	VERB
bracis-33593	13	3	these	these	DET
bracis-33593	13	4	models	model	NOUN
bracis-33593	13	5	,	,	PUNCT
bracis-33593	13	6	the	the	DET
bracis-33593	13	7	agent	agent	NOUN
bracis-33593	13	8	can	can	AUX
bracis-33593	13	9	simulate	simulate	VERB
bracis-33593	13	10	possible	possible	ADJ
bracis-33593	13	11	outcomes	outcome	NOUN
bracis-33593	13	12	of	of	ADP
bracis-33593	13	13	actions	action	NOUN
bracis-33593	13	14	without	without	ADP
bracis-33593	13	15	directly	directly	ADV
bracis-33593	13	16	interacting	interact	VERB
bracis-33593	13	17	with	with	ADP
bracis-33593	13	18	the	the	DET
bracis-33593	13	19	environment	environment	NOUN
bracis-33593	13	20	,	,	PUNCT
bracis-33593	13	21	or	or	CCONJ
bracis-33593	13	22	explore	explore	VERB
bracis-33593	13	23	a	a	DET
bracis-33593	13	24	state	state	NOUN
bracis-33593	14	1	[	[	X
bracis-33593	14	2	11	11	NUM
bracis-33593	14	3	]	]	PUNCT
bracis-33593	14	4	.	.	PUNCT
bracis-33593	15	1	this	this	PRON
bracis-33593	15	2	can	can	AUX
bracis-33593	15	3	potentially	potentially	ADV
bracis-33593	15	4	reduce	reduce	VERB
bracis-33593	15	5	the	the	DET
bracis-33593	15	6	number	number	NOUN
bracis-33593	15	7	of	of	ADP
bracis-33593	15	8	interactions	interaction	NOUN
bracis-33593	15	9	needed	need	VERB
bracis-33593	15	10	to	to	PART
bracis-33593	15	11	learn	learn	VERB
bracis-33593	15	12	an	an	DET
bracis-33593	15	13	effective	effective	ADJ
bracis-33593	15	14	policy	policy	NOUN
bracis-33593	15	15	.	.	PUNCT
bracis-33593	16	1	however	however	ADV
bracis-33593	16	2	,	,	PUNCT
bracis-33593	16	3	model	model	NOUN
bracis-33593	16	4	-	-	PUNCT
bracis-33593	16	5	based	base	VERB
bracis-33593	16	6	methods	method	NOUN
bracis-33593	16	7	may	may	AUX
bracis-33593	16	8	suffer	suffer	VERB
bracis-33593	16	9	from	from	ADP
bracis-33593	16	10	errors	error	NOUN
bracis-33593	16	11	or	or	CCONJ
bracis-33593	16	12	inaccuracies	inaccuracy	NOUN
bracis-33593	16	13	in	in	ADP
bracis-33593	16	14	the	the	DET
bracis-33593	16	15	learned	learn	VERB
bracis-33593	16	16	model	model	NOUN
bracis-33593	16	17	,	,	PUNCT
bracis-33593	16	18	leading	lead	VERB
bracis-33593	16	19	to	to	ADP
bracis-33593	16	20	sub	sub	ADJ
bracis-33593	16	21	-	-	ADJ
bracis-33593	16	22	optimal	optimal	ADJ
bracis-33593	16	23	performance	performance	NOUN
bracis-33593	16	24	[	[	X
bracis-33593	16	25	2	2	NUM
bracis-33593	16	26	]	]	PUNCT
bracis-33593	16	27	.	.	PUNCT
bracis-33593	17	1	the	the	DET
bracis-33593	17	2	concerns	concern	NOUN
bracis-33593	17	3	of	of	ADP
bracis-33593	17	4	model	model	NOUN
bracis-33593	17	5	-	-	PUNCT
bracis-33593	17	6	based	base	VERB
bracis-33593	17	7	rl	rl	X
bracis-33593	17	8	lies	lie	VERB
bracis-33593	17	9	in	in	ADP
bracis-33593	17	10	the	the	DET
bracis-33593	17	11	balance	balance	NOUN
bracis-33593	17	12	between	between	ADP
bracis-33593	17	13	the	the	DET
bracis-33593	17	14	accuracy	accuracy	NOUN
bracis-33593	17	15	of	of	ADP
bracis-33593	17	16	the	the	DET
bracis-33593	17	17	learned	learn	VERB
bracis-33593	17	18	model	model	NOUN
bracis-33593	17	19	and	and	CCONJ
bracis-33593	17	20	the	the	DET
bracis-33593	17	21	performance	performance	NOUN
bracis-33593	17	22	achieved	achieve	VERB
bracis-33593	17	23	by	by	ADP
bracis-33593	17	24	utilizing	utilize	VERB
bracis-33593	17	25	it	it	PRON
bracis-33593	17	26	to	to	PART
bracis-33593	17	27	derive	derive	VERB
bracis-33593	17	28	a	a	DET
bracis-33593	17	29	policy	policy	NOUN
bracis-33593	17	30	.	.	PUNCT
bracis-33593	18	1	by	by	ADP
bracis-33593	18	2	incorporating	incorporate	VERB
bracis-33593	18	3	learned	learn	VERB
bracis-33593	18	4	models	model	NOUN
bracis-33593	18	5	into	into	ADP
bracis-33593	18	6	the	the	DET
bracis-33593	18	7	decision	decision	NOUN
bracis-33593	18	8	-	-	PUNCT
bracis-33593	18	9	making	make	VERB
bracis-33593	18	10	process	process	NOUN
bracis-33593	18	11	,	,	PUNCT
bracis-33593	18	12	these	these	DET
bracis-33593	18	13	methods	method	NOUN
bracis-33593	18	14	can	can	AUX
bracis-33593	18	15	adapt	adapt	VERB
bracis-33593	18	16	more	more	ADV
bracis-33593	18	17	readily	readily	ADV
bracis-33593	18	18	to	to	ADP
bracis-33593	18	19	new	new	ADJ
bracis-33593	18	20	situations	situation	NOUN
bracis-33593	18	21	and	and	CCONJ
bracis-33593	18	22	exploit	exploit	VERB
bracis-33593	18	23	previously	previously	ADV
bracis-33593	18	24	acquired	acquire	VERB
bracis-33593	18	25	knowledge	knowledge	NOUN
bracis-33593	18	26	to	to	PART
bracis-33593	18	27	achieve	achieve	VERB
bracis-33593	18	28	better	well	ADJ
bracis-33593	18	29	performance	performance	NOUN
bracis-33593	18	30	.	.	PUNCT
bracis-33593	19	1	as	as	SCONJ
bracis-33593	19	2	demonstrated	demonstrate	VERB
bracis-33593	19	3	by	by	ADP
bracis-33593	19	4	young	young	PROPN
bracis-33593	19	5	et	et	PROPN
bracis-33593	19	6	al	al	PROPN
bracis-33593	19	7	.	.	PUNCT
bracis-33593	20	1	[	[	X
bracis-33593	20	2	14	14	NUM
bracis-33593	20	3	]	]	PUNCT
bracis-33593	20	4	the	the	DET
bracis-33593	20	5	benefits	benefit	NOUN
bracis-33593	20	6	of	of	ADP
bracis-33593	20	7	the	the	DET
bracis-33593	20	8	model	model	NOUN
bracis-33593	20	9	-	-	PUNCT
bracis-33593	20	10	based	base	VERB
bracis-33593	20	11	approach	approach	NOUN
bracis-33593	20	12	could	could	AUX
bracis-33593	20	13	go	go	VERB
bracis-33593	20	14	even	even	ADV
bracis-33593	20	15	further	far	ADV
bracis-33593	20	16	at	at	ADP
bracis-33593	20	17	the	the	DET
bracis-33593	20	18	point	point	NOUN
bracis-33593	20	19	of	of	ADP
bracis-33593	20	20	presenting	present	VERB
bracis-33593	20	21	better	well	ADJ
bracis-33593	20	22	solutions	solution	NOUN
bracis-33593	20	23	for	for	ADP
bracis-33593	20	24	some	some	DET
bracis-33593	20	25	problems	problem	NOUN
bracis-33593	20	26	,	,	PUNCT
bracis-33593	20	27	as	as	ADP
bracis-33593	20	28	in	in	ADP
bracis-33593	20	29	cases	case	NOUN
bracis-33593	20	30	where	where	SCONJ
bracis-33593	20	31	the	the	DET
bracis-33593	20	32	model	model	NOUN
bracis-33593	20	33	’s	’s	PART
bracis-33593	20	34	error	error	NOUN
bracis-33593	20	35	account	account	NOUN
bracis-33593	20	36	for	for	ADP
bracis-33593	20	37	an	an	DET
bracis-33593	20	38	unintended	unintended	ADJ
bracis-33593	20	39	exploration	exploration	NOUN
bracis-33593	20	40	that	that	PRON
bracis-33593	20	41	can	can	AUX
bracis-33593	20	42	have	have	VERB
bracis-33593	20	43	a	a	DET
bracis-33593	20	44	high	high	ADJ
bracis-33593	20	45	return	return	NOUN
bracis-33593	20	46	payoff	payoff	NOUN
bracis-33593	20	47	.	.	PUNCT
bracis-33593	21	1	we	we	PRON
bracis-33593	21	2	intend	intend	VERB
bracis-33593	21	3	to	to	PART
bracis-33593	21	4	delve	delve	VERB
bracis-33593	21	5	into	into	ADP
bracis-33593	21	6	the	the	DET
bracis-33593	21	7	nuanced	nuanced	ADJ
bracis-33593	21	8	advantages	advantage	NOUN
bracis-33593	21	9	of	of	ADP
bracis-33593	21	10	model	model	NOUN
bracis-33593	21	11	-	-	PUNCT
bracis-33593	21	12	based	base	VERB
bracis-33593	21	13	solutions	solution	NOUN
bracis-33593	21	14	,	,	PUNCT
bracis-33593	21	15	particularly	particularly	ADV
bracis-33593	21	16	in	in	ADP
bracis-33593	21	17	environments	environment	NOUN
bracis-33593	21	18	characterized	characterize	VERB
bracis-33593	21	19	by	by	ADP
bracis-33593	21	20	distinct	distinct	ADJ
bracis-33593	21	21	regions	region	NOUN
bracis-33593	21	22	with	with	ADP
bracis-33593	21	23	varying	vary	VERB
bracis-33593	21	24	dynamics	dynamic	NOUN
bracis-33593	21	25	.	.	PUNCT
bracis-33593	22	1	consider	consider	VERB
bracis-33593	22	2	a	a	DET
bracis-33593	22	3	scenario	scenario	NOUN
bracis-33593	22	4	where	where	SCONJ
bracis-33593	22	5	a	a	DET
bracis-33593	22	6	robot	robot	NOUN
bracis-33593	22	7	operates	operate	VERB
bracis-33593	22	8	in	in	ADP
bracis-33593	22	9	an	an	DET
bracis-33593	22	10	environment	environment	NOUN
bracis-33593	22	11	comprising	comprise	VERB
bracis-33593	22	12	two	two	NUM
bracis-33593	22	13	areas	area	NOUN
bracis-33593	22	14	with	with	ADP
bracis-33593	22	15	markedly	markedly	ADV
bracis-33593	22	16	different	different	ADJ
bracis-33593	22	17	frictional	frictional	ADJ
bracis-33593	22	18	properties	property	NOUN
bracis-33593	22	19	on	on	ADP
bracis-33593	22	20	their	their	PRON
bracis-33593	22	21	floors	floor	NOUN
bracis-33593	22	22	.	.	PUNCT
bracis-33593	23	1	in	in	ADP
bracis-33593	23	2	one	one	NUM
bracis-33593	23	3	area	area	NOUN
bracis-33593	23	4	,	,	PUNCT
bracis-33593	23	5	the	the	DET
bracis-33593	23	6	effects	effect	NOUN
bracis-33593	23	7	of	of	ADP
bracis-33593	23	8	actions	action	NOUN
bracis-33593	23	9	accounting	account	VERB
bracis-33593	23	10	for	for	ADP
bracis-33593	23	11	movement	movement	NOUN
bracis-33593	23	12	differ	differ	VERB
bracis-33593	23	13	significantly	significantly	ADV
bracis-33593	23	14	from	from	ADP
bracis-33593	23	15	those	those	PRON
bracis-33593	23	16	in	in	ADP
bracis-33593	23	17	the	the	DET
bracis-33593	23	18	other	other	ADJ
bracis-33593	23	19	.	.	PUNCT
bracis-33593	24	1	this	this	DET
bracis-33593	24	2	paper	paper	NOUN
bracis-33593	24	3	aim	aim	VERB
bracis-33593	24	4	to	to	PART
bracis-33593	24	5	explore	explore	VERB
bracis-33593	24	6	how	how	SCONJ
bracis-33593	24	7	assumptions	assumption	NOUN
bracis-33593	24	8	like	like	ADP
bracis-33593	24	9	the	the	DET
bracis-33593	24	10	existence	existence	NOUN
bracis-33593	24	11	of	of	ADP
bracis-33593	24	12	distinct	distinct	ADJ
bracis-33593	24	13	regions	region	NOUN
bracis-33593	24	14	in	in	ADP
bracis-33593	24	15	model	model	NOUN
bracis-33593	24	16	-	-	PUNCT
bracis-33593	24	17	based	base	VERB
bracis-33593	24	18	solutions	solution	NOUN
bracis-33593	24	19	can	can	AUX
bracis-33593	24	20	enhance	enhance	VERB
bracis-33593	24	21	its	its	PRON
bracis-33593	24	22	capabilities	capability	NOUN
bracis-33593	24	23	of	of	ADP
bracis-33593	24	24	planning	plan	VERB
bracis-33593	24	25	.	.	PUNCT
bracis-33593	25	1	constructing	construct	VERB
bracis-33593	25	2	multiple	multiple	ADJ
bracis-33593	25	3	models	model	NOUN
bracis-33593	25	4	as	as	SCONJ
bracis-33593	25	5	done	do	VERB
bracis-33593	25	6	by	by	ADP
bracis-33593	25	7	doya	doya	NOUN
bracis-33593	25	8	et	et	PROPN
bracis-33593	25	9	al	al	PROPN
bracis-33593	25	10	.	.	PUNCT
bracis-33593	26	1	[	[	X
bracis-33593	26	2	3	3	X
bracis-33593	26	3	]	]	PUNCT
bracis-33593	26	4	tailored	tailor	VERB
bracis-33593	26	5	to	to	ADP
bracis-33593	26	6	individual	individual	ADJ
bracis-33593	26	7	regions	region	NOUN
bracis-33593	26	8	exhibiting	exhibit	VERB
bracis-33593	26	9	stable	stable	ADJ
bracis-33593	26	10	dynamics	dynamic	NOUN
bracis-33593	26	11	,	,	PUNCT
bracis-33593	26	12	rather	rather	ADV
bracis-33593	26	13	than	than	ADP
bracis-33593	26	14	attempting	attempt	VERB
bracis-33593	26	15	to	to	PART
bracis-33593	26	16	capture	capture	VERB
bracis-33593	26	17	the	the	DET
bracis-33593	26	18	complex	complex	ADJ
bracis-33593	26	19	dynamics	dynamic	NOUN
bracis-33593	26	20	of	of	ADP
bracis-33593	26	21	both	both	DET
bracis-33593	26	22	regions	region	NOUN
bracis-33593	26	23	within	within	ADP
bracis-33593	26	24	a	a	DET
bracis-33593	26	25	single	single	ADJ
bracis-33593	26	26	model	model	NOUN
bracis-33593	26	27	could	could	AUX
bracis-33593	26	28	yield	yield	VERB
bracis-33593	26	29	better	well	ADJ
bracis-33593	26	30	results	result	NOUN
bracis-33593	26	31	.	.	PUNCT
bracis-33593	27	1	we	we	PRON
bracis-33593	27	2	propose	propose	VERB
bracis-33593	27	3	an	an	DET
bracis-33593	27	4	approach	approach	NOUN
bracis-33593	27	5	wherein	wherein	SCONJ
bracis-33593	27	6	distinct	distinct	ADJ
bracis-33593	27	7	models	model	NOUN
bracis-33593	27	8	are	be	AUX
bracis-33593	27	9	built	build	VERB
bracis-33593	27	10	for	for	ADP
bracis-33593	27	11	each	each	DET
bracis-33593	27	12	identifiable	identifiable	ADJ
bracis-33593	27	13	region	region	NOUN
bracis-33593	27	14	,	,	PUNCT
bracis-33593	27	15	and	and	CCONJ
bracis-33593	27	16	selected	select	VERB
bracis-33593	27	17	for	for	ADP
bracis-33593	27	18	use	use	NOUN
bracis-33593	27	19	based	base	VERB
bracis-33593	27	20	on	on	ADP
bracis-33593	27	21	agent	agent	NOUN
bracis-33593	27	22	past	past	ADP
bracis-33593	27	23	experience	experience	NOUN
bracis-33593	27	24	.	.	PUNCT
bracis-33593	28	1	the	the	DET
bracis-33593	28	2	learned	learn	VERB
bracis-33593	28	3	models	model	NOUN
bracis-33593	28	4	are	be	AUX
bracis-33593	28	5	leveraged	leverage	VERB
bracis-33593	28	6	for	for	ADP
bracis-33593	28	7	plan	plan	NOUN
bracis-33593	28	8	generation	generation	NOUN
bracis-33593	28	9	through	through	ADP
bracis-33593	28	10	the	the	DET
bracis-33593	28	11	utilization	utilization	NOUN
bracis-33593	28	12	of	of	ADP
bracis-33593	28	13	a	a	DET
bracis-33593	28	14	rapidly	rapidly	ADV
bracis-33593	28	15	exploring	explore	VERB
bracis-33593	28	16	random	random	ADJ
bracis-33593	28	17	tree	tree	NOUN
bracis-33593	28	18	(	(	PUNCT
bracis-33593	28	19	rrt	rrt	NOUN
bracis-33593	28	20	)	)	PUNCT
bracis-33593	28	21	based	base	VERB
bracis-33593	28	22	algorithm	algorithm	NOUN
bracis-33593	28	23	.	.	PUNCT
bracis-33593	29	1	our	our	PRON
bracis-33593	29	2	planning	planning	NOUN
bracis-33593	29	3	strategy	strategy	NOUN
bracis-33593	29	4	draws	draw	VERB
bracis-33593	29	5	inspiration	inspiration	NOUN
bracis-33593	29	6	from	from	ADP
bracis-33593	29	7	the	the	DET
bracis-33593	29	8	work	work	NOUN
bracis-33593	29	9	of	of	ADP
bracis-33593	29	10	lin	lin	PROPN
bracis-33593	29	11	 	 	SPACE
bracis-33593	29	12	and	and	CCONJ
bracis-33593	29	13	zhang	zhang	PROPN
bracis-33593	30	1	[	[	X
bracis-33593	30	2	9	9	NUM
bracis-33593	30	3	]	]	PUNCT
bracis-33593	30	4	,	,	PUNCT
bracis-33593	30	5	which	which	PRON
bracis-33593	30	6	introduced	introduce	VERB
bracis-33593	30	7	the	the	DET
bracis-33593	30	8	integration	integration	NOUN
bracis-33593	30	9	of	of	ADP
bracis-33593	30	10	rrt	rrt	NOUN
bracis-33593	30	11	with	with	ADP
bracis-33593	30	12	reinforcement	reinforcement	NOUN
bracis-33593	30	13	learning	learning	NOUN
bracis-33593	30	14	for	for	ADP
bracis-33593	30	15	planning	planning	NOUN
bracis-33593	30	16	tasks	task	NOUN
bracis-33593	30	17	.	.	PUNCT
bracis-33593	31	1	by	by	ADP
bracis-33593	31	2	adopting	adopt	VERB
bracis-33593	31	3	this	this	DET
bracis-33593	31	4	multi	multi	ADJ
bracis-33593	31	5	-	-	ADJ
bracis-33593	31	6	model	model	ADJ
bracis-33593	31	7	approach	approach	NOUN
bracis-33593	31	8	,	,	PUNCT
bracis-33593	31	9	we	we	PRON
bracis-33593	31	10	aim	aim	VERB
bracis-33593	31	11	to	to	PART
bracis-33593	31	12	investigate	investigate	VERB
bracis-33593	31	13	its	its	PRON
bracis-33593	31	14	efficacy	efficacy	NOUN
bracis-33593	31	15	in	in	ADP
bracis-33593	31	16	accurately	accurately	ADV
bracis-33593	31	17	representing	represent	VERB
bracis-33593	31	18	the	the	DET
bracis-33593	31	19	environment	environment	NOUN
bracis-33593	31	20	’s	’s	PART
bracis-33593	31	21	dynamics	dynamic	NOUN
bracis-33593	31	22	and	and	CCONJ
bracis-33593	31	23	its	its	PRON
bracis-33593	31	24	impact	impact	NOUN
bracis-33593	31	25	on	on	ADP
bracis-33593	31	26	the	the	DET
bracis-33593	31	27	agent	agent	NOUN
bracis-33593	31	28	’s	’s	PART
bracis-33593	31	29	decision	decision	NOUN
bracis-33593	31	30	-	-	PUNCT
bracis-33593	31	31	making	make	VERB
bracis-33593	31	32	process	process	NOUN
bracis-33593	31	33	.	.	PUNCT
bracis-33593	32	1	we	we	PRON
bracis-33593	32	2	hypothesize	hypothesize	VERB
bracis-33593	32	3	that	that	SCONJ
bracis-33593	32	4	by	by	ADP
bracis-33593	32	5	aligning	align	VERB
bracis-33593	32	6	models	model	NOUN
bracis-33593	32	7	with	with	ADP
bracis-33593	32	8	specific	specific	ADJ
bracis-33593	32	9	regions	region	NOUN
bracis-33593	32	10	,	,	PUNCT
bracis-33593	32	11	the	the	DET
bracis-33593	32	12	agent	agent	NOUN
bracis-33593	32	13	can	can	AUX
bracis-33593	32	14	better	well	ADV
bracis-33593	32	15	understand	understand	VERB
bracis-33593	32	16	and	and	CCONJ
bracis-33593	32	17	adapt	adapt	VERB
bracis-33593	32	18	to	to	ADP
bracis-33593	32	19	the	the	DET
bracis-33593	32	20	unique	unique	ADJ
bracis-33593	32	21	characteristics	characteristic	NOUN
bracis-33593	32	22	of	of	ADP
bracis-33593	32	23	each	each	DET
bracis-33593	32	24	area	area	NOUN
bracis-33593	32	25	,	,	PUNCT
bracis-33593	32	26	ultimately	ultimately	ADV
bracis-33593	32	27	leading	lead	VERB
bracis-33593	32	28	to	to	ADP
bracis-33593	32	29	improved	improved	ADJ
bracis-33593	32	30	performance	performance	NOUN
bracis-33593	32	31	and	and	CCONJ
bracis-33593	32	32	efficiency	efficiency	NOUN
bracis-33593	32	33	in	in	ADP
bracis-33593	32	34	task	task	NOUN
bracis-33593	32	35	execution	execution	NOUN
bracis-33593	32	36	.	.	PUNCT
bracis-33593	33	1	through	through	ADP
bracis-33593	33	2	empirical	empirical	ADJ
bracis-33593	33	3	evaluation	evaluation	NOUN
bracis-33593	33	4	and	and	CCONJ
bracis-33593	33	5	analysis	analysis	NOUN
bracis-33593	33	6	,	,	PUNCT
bracis-33593	33	7	we	we	PRON
bracis-33593	33	8	aim	aim	VERB
bracis-33593	33	9	to	to	PART
bracis-33593	33	10	provide	provide	VERB
bracis-33593	33	11	insights	insight	NOUN
bracis-33593	33	12	into	into	ADP
bracis-33593	33	13	the	the	DET
bracis-33593	33	14	comparative	comparative	ADJ
bracis-33593	33	15	advantages	advantage	NOUN
bracis-33593	33	16	and	and	CCONJ
bracis-33593	33	17	limitations	limitation	NOUN
bracis-33593	33	18	of	of	ADP
bracis-33593	33	19	both	both	DET
bracis-33593	33	20	traditional	traditional	ADJ
bracis-33593	33	21	single	single	ADJ
bracis-33593	33	22	-	-	PUNCT
bracis-33593	33	23	model	model	NOUN
bracis-33593	33	24	and	and	CCONJ
bracis-33593	33	25	proposed	propose	VERB
bracis-33593	33	26	segmented	segment	VERB
bracis-33593	33	27	model	model	NOUN
bracis-33593	33	28	approaches	approach	NOUN
bracis-33593	33	29	in	in	ADP
bracis-33593	33	30	addressing	address	VERB
bracis-33593	33	31	challenges	challenge	NOUN
bracis-33593	33	32	posed	pose	VERB
bracis-33593	33	33	by	by	ADP
bracis-33593	33	34	environments	environment	NOUN
bracis-33593	33	35	with	with	ADP
bracis-33593	33	36	heterogeneous	heterogeneous	ADJ
bracis-33593	33	37	dynamics	dynamic	NOUN
bracis-33593	33	38	.	.	PUNCT
bracis-33593	34	1	our	our	PRON
bracis-33593	34	2	findings	finding	NOUN
bracis-33593	34	3	aim	aim	VERB
bracis-33593	34	4	to	to	PART
bracis-33593	34	5	contribute	contribute	VERB
bracis-33593	34	6	to	to	ADP
bracis-33593	34	7	a	a	DET
bracis-33593	34	8	deeper	deep	ADJ
bracis-33593	34	9	understanding	understanding	NOUN
bracis-33593	34	10	of	of	ADP
bracis-33593	34	11	the	the	DET
bracis-33593	34	12	applicability	applicability	NOUN
bracis-33593	34	13	of	of	ADP
bracis-33593	34	14	model	model	NOUN
bracis-33593	34	15	-	-	PUNCT
bracis-33593	34	16	based	base	VERB
bracis-33593	34	17	solutions	solution	NOUN
bracis-33593	34	18	in	in	ADP
bracis-33593	34	19	real	real	ADJ
bracis-33593	34	20	-	-	PUNCT
bracis-33593	34	21	world	world	NOUN
bracis-33593	34	22	scenarios	scenario	NOUN
bracis-33593	34	23	characterized	characterize	VERB
bracis-33593	34	24	by	by	ADP
bracis-33593	34	25	spatially	spatially	ADV
bracis-33593	34	26	varying	vary	VERB
bracis-33593	34	27	dynamics	dynamic	NOUN
bracis-33593	34	28	.	.	PUNCT
bracis-33593	35	1	in	in	ADP
bracis-33593	35	2	sect	sect	NOUN
bracis-33593	35	3	.	.	PUNCT
bracis-33593	35	4	 	 	SPACE
bracis-33593	35	5	2	2	NUM
bracis-33593	35	6	of	of	ADP
bracis-33593	35	7	our	our	PRON
bracis-33593	35	8	paper	paper	NOUN
bracis-33593	35	9	,	,	PUNCT
bracis-33593	35	10	we	we	PRON
bracis-33593	35	11	delve	delve	VERB
bracis-33593	35	12	into	into	ADP
bracis-33593	35	13	related	related	ADJ
bracis-33593	35	14	works	work	NOUN
bracis-33593	35	15	that	that	PRON
bracis-33593	35	16	have	have	AUX
bracis-33593	35	17	explored	explore	VERB
bracis-33593	35	18	the	the	DET
bracis-33593	35	19	concept	concept	NOUN
bracis-33593	35	20	of	of	ADP
bracis-33593	35	21	leveraging	leverage	VERB
bracis-33593	35	22	multiple	multiple	ADJ
bracis-33593	35	23	models	model	NOUN
bracis-33593	35	24	to	to	PART
bracis-33593	35	25	enhance	enhance	VERB
bracis-33593	35	26	generalization	generalization	NOUN
bracis-33593	35	27	in	in	ADP
bracis-33593	35	28	rl	rl	PROPN
bracis-33593	35	29	.	.	PROPN
bracis-33593	35	30	section	section	NOUN
bracis-33593	35	31	 	 	SPACE
bracis-33593	35	32	3	3	NUM
bracis-33593	35	33	formalizes	formalize	VERB
bracis-33593	35	34	the	the	DET
bracis-33593	35	35	segmented	segment	VERB
bracis-33593	35	36	environment	environment	NOUN
bracis-33593	35	37	problem	problem	NOUN
bracis-33593	35	38	in	in	ADP
bracis-33593	35	39	terms	term	NOUN
bracis-33593	35	40	of	of	ADP
bracis-33593	35	41	a	a	DET
bracis-33593	35	42	markov	markov	NOUN
bracis-33593	35	43	decision	decision	NOUN
bracis-33593	35	44	problem	problem	NOUN
bracis-33593	35	45	.	.	PUNCT
bracis-33593	36	1	section	section	NOUN
bracis-33593	36	2	 	 	SPACE
bracis-33593	36	3	4	4	NUM
bracis-33593	36	4	introduces	introduce	VERB
bracis-33593	36	5	our	our	PRON
bracis-33593	36	6	approach	approach	NOUN
bracis-33593	36	7	based	base	VERB
bracis-33593	36	8	on	on	ADP
bracis-33593	36	9	segmented	segment	VERB
bracis-33593	36	10	models	model	NOUN
bracis-33593	36	11	problem	problem	VERB
bracis-33593	36	12	that	that	PRON
bracis-33593	36	13	forms	form	VERB
bracis-33593	36	14	the	the	DET
bracis-33593	36	15	basis	basis	NOUN
bracis-33593	36	16	of	of	ADP
bracis-33593	36	17	our	our	PRON
bracis-33593	36	18	comparative	comparative	ADJ
bracis-33593	36	19	analysis	analysis	NOUN
bracis-33593	36	20	.	.	PUNCT
bracis-33593	37	1	moving	move	VERB
bracis-33593	37	2	on	on	ADP
bracis-33593	37	3	to	to	PART
bracis-33593	37	4	sect	sect	VERB
bracis-33593	37	5	.	.	PUNCT
bracis-33593	37	6	 	 	SPACE
bracis-33593	38	1	5	5	NUM
bracis-33593	38	2	,	,	PUNCT
bracis-33593	38	3	we	we	PRON
bracis-33593	38	4	provide	provide	VERB
bracis-33593	38	5	a	a	DET
bracis-33593	38	6	concise	concise	ADJ
bracis-33593	38	7	description	description	NOUN
bracis-33593	38	8	of	of	ADP
bracis-33593	38	9	the	the	DET
bracis-33593	38	10	simple	simple	ADJ
bracis-33593	38	11	environment	environment	NOUN
bracis-33593	38	12	utilized	utilize	VERB
bracis-33593	38	13	for	for	ADP
bracis-33593	38	14	collecting	collect	VERB
bracis-33593	38	15	empirical	empirical	ADJ
bracis-33593	38	16	results	result	NOUN
bracis-33593	38	17	.	.	PUNCT
bracis-33593	39	1	finally	finally	ADV
bracis-33593	39	2	,	,	PUNCT
bracis-33593	39	3	in	in	ADP
bracis-33593	39	4	sect	sect	NOUN
bracis-33593	39	5	.	.	PUNCT
bracis-33593	39	6	 	 	SPACE
bracis-33593	39	7	6	6	NUM
bracis-33593	39	8	,	,	PUNCT
bracis-33593	39	9	we	we	PRON
bracis-33593	39	10	present	present	VERB
bracis-33593	39	11	the	the	DET
bracis-33593	39	12	results	result	NOUN
bracis-33593	39	13	of	of	ADP
bracis-33593	39	14	our	our	PRON
bracis-33593	39	15	empirical	empirical	ADJ
bracis-33593	39	16	evaluations	evaluation	NOUN
bracis-33593	39	17	,	,	PUNCT
bracis-33593	39	18	detailing	detail	VERB
bracis-33593	39	19	the	the	DET
bracis-33593	39	20	performance	performance	NOUN
bracis-33593	39	21	metrics	metric	NOUN
bracis-33593	39	22	and	and	CCONJ
bracis-33593	39	23	outcomes	outcome	NOUN
bracis-33593	39	24	observed	observe	VERB
bracis-33593	39	25	across	across	ADP
bracis-33593	39	26	different	different	ADJ
bracis-33593	39	27	experimental	experimental	ADJ
bracis-33593	39	28	conditions	condition	NOUN
bracis-33593	39	29	.	.	PUNCT
bracis-33593	40	1	2	2	NUM
bracis-33593	40	2	related	relate	VERB
bracis-33593	40	3	works	work	NOUN
bracis-33593	40	4	several	several	ADJ
bracis-33593	40	5	existing	exist	VERB
bracis-33593	40	6	works	work	NOUN
bracis-33593	40	7	have	have	AUX
bracis-33593	40	8	explored	explore	VERB
bracis-33593	40	9	the	the	DET
bracis-33593	40	10	concept	concept	NOUN
bracis-33593	40	11	of	of	ADP
bracis-33593	40	12	constructing	construct	VERB
bracis-33593	40	13	a	a	DET
bracis-33593	40	14	segmented	segmented	ADJ
bracis-33593	40	15	model	model	NOUN
bracis-33593	40	16	of	of	ADP
bracis-33593	40	17	the	the	DET
bracis-33593	40	18	environment	environment	NOUN
bracis-33593	40	19	’s	’s	PART
bracis-33593	40	20	dynamics	dynamic	NOUN
bracis-33593	40	21	,	,	PUNCT
bracis-33593	40	22	aiming	aim	VERB
bracis-33593	40	23	to	to	PART
bracis-33593	40	24	create	create	VERB
bracis-33593	40	25	a	a	DET
bracis-33593	40	26	comprehensive	comprehensive	ADJ
bracis-33593	40	27	understanding	understanding	NOUN
bracis-33593	40	28	of	of	ADP
bracis-33593	40	29	the	the	DET
bracis-33593	40	30	environment	environment	NOUN
bracis-33593	40	31	.	.	PUNCT
bracis-33593	41	1	one	one	NUM
bracis-33593	41	2	notable	notable	ADJ
bracis-33593	41	3	example	example	NOUN
bracis-33593	41	4	is	be	AUX
bracis-33593	41	5	the	the	DET
bracis-33593	41	6	work	work	NOUN
bracis-33593	41	7	by	by	ADP
bracis-33593	41	8	doya	doya	NOUN
bracis-33593	41	9	et	et	PROPN
bracis-33593	41	10	al	al	PROPN
bracis-33593	41	11	.	.	PUNCT
bracis-33593	42	1	[	[	X
bracis-33593	42	2	3	3	NUM
bracis-33593	42	3	]	]	PUNCT
bracis-33593	42	4	,	,	PUNCT
bracis-33593	42	5	which	which	PRON
bracis-33593	42	6	introduces	introduce	VERB
bracis-33593	42	7	a	a	DET
bracis-33593	42	8	concept	concept	NOUN
bracis-33593	42	9	akin	akin	ADJ
bracis-33593	42	10	to	to	ADP
bracis-33593	42	11	a	a	DET
bracis-33593	42	12	mixture	mixture	NOUN
bracis-33593	42	13	of	of	ADP
bracis-33593	42	14	specialists	specialist	NOUN
bracis-33593	42	15	.	.	PUNCT
bracis-33593	43	1	in	in	ADP
bracis-33593	43	2	this	this	DET
bracis-33593	43	3	approach	approach	NOUN
bracis-33593	43	4	,	,	PUNCT
bracis-33593	43	5	the	the	DET
bracis-33593	43	6	environment	environment	NOUN
bracis-33593	43	7	’s	’s	PART
bracis-33593	43	8	dynamics	dynamic	NOUN
bracis-33593	43	9	are	be	AUX
bracis-33593	43	10	modeled	model	VERB
bracis-33593	43	11	by	by	ADP
bracis-33593	43	12	a	a	DET
bracis-33593	43	13	collection	collection	NOUN
bracis-33593	43	14	of	of	ADP
bracis-33593	43	15	smaller	small	ADJ
bracis-33593	43	16	models	model	NOUN
bracis-33593	43	17	,	,	PUNCT
bracis-33593	43	18	each	each	PRON
bracis-33593	43	19	specialized	specialize	VERB
bracis-33593	43	20	for	for	ADP
bracis-33593	43	21	specific	specific	ADJ
bracis-33593	43	22	states	state	NOUN
bracis-33593	43	23	or	or	CCONJ
bracis-33593	43	24	moments	moment	NOUN
bracis-33593	43	25	where	where	SCONJ
bracis-33593	43	26	distinct	distinct	ADJ
bracis-33593	43	27	behaviors	behavior	NOUN
bracis-33593	43	28	are	be	AUX
bracis-33593	43	29	observed	observe	VERB
bracis-33593	43	30	.	.	PUNCT
bracis-33593	44	1	the	the	DET
bracis-33593	44	2	fundamental	fundamental	ADJ
bracis-33593	44	3	difference	difference	NOUN
bracis-33593	44	4	between	between	ADP
bracis-33593	44	5	observed	observed	ADJ
bracis-33593	44	6	approach	approach	NOUN
bracis-33593	44	7	[	[	X
bracis-33593	44	8	3	3	X
bracis-33593	44	9	]	]	PUNCT
bracis-33593	44	10	and	and	CCONJ
bracis-33593	44	11	our	our	PRON
bracis-33593	44	12	methodology	methodology	NOUN
bracis-33593	44	13	lies	lie	VERB
bracis-33593	44	14	in	in	ADP
bracis-33593	44	15	the	the	DET
bracis-33593	44	16	mechanism	mechanism	NOUN
bracis-33593	44	17	for	for	ADP
bracis-33593	44	18	combining	combine	VERB
bracis-33593	44	19	the	the	DET
bracis-33593	44	20	outputs	output	NOUN
bracis-33593	44	21	of	of	ADP
bracis-33593	44	22	these	these	DET
bracis-33593	44	23	specialized	specialized	ADJ
bracis-33593	44	24	models	model	NOUN
bracis-33593	44	25	.	.	PUNCT
bracis-33593	45	1	in	in	ADP
bracis-33593	45	2	multiple	multiple	ADJ
bracis-33593	45	3	model	model	NOUN
bracis-33593	45	4	based	base	VERB
bracis-33593	45	5	reinforcemnt	reinforcemnt	PROPN
bracis-33593	45	6	learning	learn	VERB
bracis-33593	45	7	[	[	X
bracis-33593	45	8	3	3	NUM
bracis-33593	45	9	]	]	PUNCT
bracis-33593	45	10	,	,	PUNCT
bracis-33593	45	11	the	the	DET
bracis-33593	45	12	specialists	specialist	NOUN
bracis-33593	45	13	contribute	contribute	VERB
bracis-33593	45	14	to	to	ADP
bracis-33593	45	15	the	the	DET
bracis-33593	45	16	overall	overall	ADJ
bracis-33593	45	17	model	model	NOUN
bracis-33593	45	18	through	through	ADP
bracis-33593	45	19	a	a	DET
bracis-33593	45	20	continuous	continuous	ADJ
bracis-33593	45	21	responsibility	responsibility	NOUN
bracis-33593	45	22	signal	signal	NOUN
bracis-33593	45	23	,	,	PUNCT
bracis-33593	45	24	allowing	allow	VERB
bracis-33593	45	25	for	for	ADP
bracis-33593	45	26	a	a	DET
bracis-33593	45	27	potentially	potentially	ADV
bracis-33593	45	28	blended	blend	VERB
bracis-33593	45	29	output	output	NOUN
bracis-33593	45	30	that	that	PRON
bracis-33593	45	31	may	may	AUX
bracis-33593	45	32	not	not	PART
bracis-33593	45	33	be	be	AUX
bracis-33593	45	34	directly	directly	ADV
bracis-33593	45	35	attributable	attributable	ADJ
bracis-33593	45	36	to	to	ADP
bracis-33593	45	37	any	any	DET
bracis-33593	45	38	individual	individual	ADJ
bracis-33593	45	39	specialist	specialist	NOUN
bracis-33593	45	40	.	.	PUNCT
bracis-33593	46	1	in	in	ADP
bracis-33593	46	2	contrast	contrast	NOUN
bracis-33593	46	3	,	,	PUNCT
bracis-33593	46	4	our	our	PRON
bracis-33593	46	5	approach	approach	NOUN
bracis-33593	46	6	adopts	adopt	VERB
bracis-33593	46	7	a	a	DET
bracis-33593	46	8	structured	structured	ADJ
bracis-33593	46	9	mechanism	mechanism	NOUN
bracis-33593	46	10	where	where	SCONJ
bracis-33593	46	11	the	the	DET
bracis-33593	46	12	output	output	NOUN
bracis-33593	46	13	is	be	AUX
bracis-33593	46	14	determined	determine	VERB
bracis-33593	46	15	by	by	ADP
bracis-33593	46	16	the	the	DET
bracis-33593	46	17	specialist	specialist	NOUN
bracis-33593	46	18	that	that	PRON
bracis-33593	46	19	maximizes	maximize	VERB
bracis-33593	46	20	its	its	PRON
bracis-33593	46	21	signal	signal	NOUN
bracis-33593	46	22	.	.	PUNCT
bracis-33593	47	1	this	this	PRON
bracis-33593	47	2	ensures	ensure	VERB
bracis-33593	47	3	that	that	SCONJ
bracis-33593	47	4	the	the	DET
bracis-33593	47	5	output	output	NOUN
bracis-33593	47	6	is	be	AUX
bracis-33593	47	7	directly	directly	ADV
bracis-33593	47	8	linked	link	VERB
bracis-33593	47	9	to	to	ADP
bracis-33593	47	10	the	the	DET
bracis-33593	47	11	most	most	ADV
bracis-33593	47	12	relevant	relevant	ADJ
bracis-33593	47	13	specialist	specialist	NOUN
bracis-33593	47	14	,	,	PUNCT
bracis-33593	47	15	facilitating	facilitate	VERB
bracis-33593	47	16	clearer	clear	ADJ
bracis-33593	47	17	interpretation	interpretation	NOUN
bracis-33593	47	18	and	and	CCONJ
bracis-33593	47	19	utilization	utilization	NOUN
bracis-33593	47	20	of	of	ADP
bracis-33593	47	21	the	the	DET
bracis-33593	47	22	model	model	NOUN
bracis-33593	47	23	’s	’s	PART
bracis-33593	47	24	predictions	prediction	NOUN
bracis-33593	47	25	.	.	PUNCT
bracis-33593	48	1	an	an	DET
bracis-33593	48	2	approach	approach	NOUN
bracis-33593	48	3	was	be	AUX
bracis-33593	48	4	proposed	propose	VERB
bracis-33593	48	5	[	[	X
bracis-33593	48	6	5	5	NUM
bracis-33593	48	7	]	]	PUNCT
bracis-33593	48	8	,	,	PUNCT
bracis-33593	48	9	which	which	PRON
bracis-33593	48	10	introduces	introduce	VERB
bracis-33593	48	11	a	a	DET
bracis-33593	48	12	novel	novel	ADJ
bracis-33593	48	13	representation	representation	NOUN
bracis-33593	48	14	of	of	ADP
bracis-33593	48	15	the	the	DET
bracis-33593	48	16	problem	problem	NOUN
bracis-33593	48	17	by	by	ADP
bracis-33593	48	18	augmenting	augment	VERB
bracis-33593	48	19	the	the	DET
bracis-33593	48	20	traditional	traditional	ADJ
bracis-33593	48	21	markov	markov	NOUN
bracis-33593	48	22	decision	decision	NOUN
bracis-33593	48	23	process	process	NOUN
bracis-33593	48	24	(	(	PUNCT
bracis-33593	48	25	mdp	mdp	NOUN
bracis-33593	48	26	)	)	PUNCT
bracis-33593	48	27	with	with	ADP
bracis-33593	48	28	contextual	contextual	ADJ
bracis-33593	48	29	information	information	NOUN
bracis-33593	48	30	.	.	PUNCT
bracis-33593	49	1	this	this	DET
bracis-33593	49	2	augmentation	augmentation	NOUN
bracis-33593	49	3	enables	enable	VERB
bracis-33593	49	4	the	the	DET
bracis-33593	49	5	inference	inference	NOUN
bracis-33593	49	6	of	of	ADP
bracis-33593	49	7	different	different	ADJ
bracis-33593	49	8	underlying	underlie	VERB
bracis-33593	49	9	models	model	NOUN
bracis-33593	49	10	based	base	VERB
bracis-33593	49	11	on	on	ADP
bracis-33593	49	12	the	the	DET
bracis-33593	49	13	context	context	NOUN
bracis-33593	49	14	provided	provide	VERB
bracis-33593	49	15	,	,	PUNCT
bracis-33593	49	16	offering	offer	VERB
bracis-33593	49	17	a	a	DET
bracis-33593	49	18	flexible	flexible	ADJ
bracis-33593	49	19	framework	framework	NOUN
bracis-33593	49	20	for	for	ADP
bracis-33593	49	21	adapting	adapt	VERB
bracis-33593	49	22	to	to	PART
bracis-33593	49	23	diverse	diverse	VERB
bracis-33593	49	24	environmental	environmental	ADJ
bracis-33593	49	25	dynamics	dynamic	NOUN
bracis-33593	49	26	.	.	PUNCT
bracis-33593	50	1	one	one	NUM
bracis-33593	50	2	notable	notable	ADJ
bracis-33593	50	3	aspect	aspect	NOUN
bracis-33593	50	4	of	of	ADP
bracis-33593	50	5	this	this	DET
bracis-33593	50	6	work	work	NOUN
bracis-33593	50	7	is	be	AUX
bracis-33593	50	8	the	the	DET
bracis-33593	50	9	similarity	similarity	NOUN
bracis-33593	50	10	in	in	ADP
bracis-33593	50	11	notation	notation	NOUN
bracis-33593	50	12	used	use	VERB
bracis-33593	50	13	for	for	ADP
bracis-33593	50	14	representing	represent	VERB
bracis-33593	50	15	context	context	NOUN
bracis-33593	50	16	,	,	PUNCT
bracis-33593	50	17	which	which	PRON
bracis-33593	50	18	bears	bear	VERB
bracis-33593	50	19	resemblance	resemblance	NOUN
bracis-33593	50	20	to	to	ADP
bracis-33593	50	21	the	the	DET
bracis-33593	50	22	representation	representation	NOUN
bracis-33593	50	23	of	of	ADP
bracis-33593	50	24	our	our	PRON
bracis-33593	50	25	models	model	NOUN
bracis-33593	50	26	.	.	PUNCT
bracis-33593	51	1	however	however	ADV
bracis-33593	51	2	,	,	PUNCT
bracis-33593	51	3	a	a	DET
bracis-33593	51	4	significant	significant	ADJ
bracis-33593	51	5	distinction	distinction	NOUN
bracis-33593	51	6	lies	lie	VERB
bracis-33593	51	7	in	in	ADP
bracis-33593	51	8	the	the	DET
bracis-33593	51	9	treatment	treatment	NOUN
bracis-33593	51	10	of	of	ADP
bracis-33593	51	11	context	context	NOUN
bracis-33593	51	12	over	over	ADP
bracis-33593	51	13	the	the	DET
bracis-33593	51	14	course	course	NOUN
bracis-33593	51	15	of	of	ADP
bracis-33593	51	16	a	a	DET
bracis-33593	51	17	trajectory	trajectory	NOUN
bracis-33593	51	18	.	.	PUNCT
bracis-33593	52	1	unlike	unlike	ADP
bracis-33593	52	2	our	our	PRON
bracis-33593	52	3	approach	approach	NOUN
bracis-33593	52	4	,	,	PUNCT
bracis-33593	52	5	which	which	PRON
bracis-33593	52	6	allows	allow	VERB
bracis-33593	52	7	for	for	ADP
bracis-33593	52	8	the	the	DET
bracis-33593	52	9	possibility	possibility	NOUN
bracis-33593	52	10	of	of	ADP
bracis-33593	52	11	encountering	encounter	VERB
bracis-33593	52	12	multiple	multiple	ADJ
bracis-33593	52	13	contexts	context	NOUN
bracis-33593	52	14	within	within	ADP
bracis-33593	52	15	the	the	DET
bracis-33593	52	16	same	same	ADJ
bracis-33593	52	17	episode	episode	NOUN
bracis-33593	52	18	,	,	PUNCT
bracis-33593	52	19	the	the	DET
bracis-33593	52	20	contextual	contextual	ADJ
bracis-33593	52	21	mdp	mdp	NOUN
bracis-33593	52	22	framework	framework	NOUN
bracis-33593	52	23	[	[	X
bracis-33593	52	24	5	5	NUM
bracis-33593	52	25	]	]	PUNCT
bracis-33593	52	26	assumes	assume	VERB
bracis-33593	52	27	a	a	DET
bracis-33593	52	28	fixed	fix	VERB
bracis-33593	52	29	context	context	NOUN
bracis-33593	52	30	for	for	ADP
bracis-33593	52	31	the	the	DET
bracis-33593	52	32	entire	entire	ADJ
bracis-33593	52	33	trajectory	trajectory	NOUN
bracis-33593	52	34	.	.	PUNCT
bracis-33593	53	1	the	the	DET
bracis-33593	53	2	idea	idea	NOUN
bracis-33593	53	3	of	of	ADP
bracis-33593	53	4	dynamically	dynamically	ADV
bracis-33593	53	5	identifying	identify	VERB
bracis-33593	53	6	the	the	DET
bracis-33593	53	7	context	context	NOUN
bracis-33593	53	8	have	have	AUX
bracis-33593	53	9	been	be	AUX
bracis-33593	53	10	addressed	address	VERB
bracis-33593	53	11	[	[	X
bracis-33593	53	12	12	12	NUM
bracis-33593	53	13	]	]	PUNCT
bracis-33593	53	14	with	with	ADP
bracis-33593	53	15	a	a	DET
bracis-33593	53	16	solution	solution	NOUN
bracis-33593	53	17	capable	capable	ADJ
bracis-33593	53	18	of	of	ADP
bracis-33593	53	19	learning	learn	VERB
bracis-33593	53	20	a	a	DET
bracis-33593	53	21	new	new	ADJ
bracis-33593	53	22	local	local	ADJ
bracis-33593	53	23	model	model	NOUN
bracis-33593	53	24	whenever	whenever	SCONJ
bracis-33593	53	25	it	it	PRON
bracis-33593	53	26	detects	detect	VERB
bracis-33593	53	27	a	a	DET
bracis-33593	53	28	context	context	NOUN
bracis-33593	53	29	not	not	PART
bracis-33593	53	30	yet	yet	ADV
bracis-33593	53	31	experienced	experience	VERB
bracis-33593	53	32	.	.	PUNCT
bracis-33593	54	1	also	also	ADV
bracis-33593	54	2	using	use	VERB
bracis-33593	54	3	as	as	ADP
bracis-33593	54	4	a	a	DET
bracis-33593	54	5	basis	basis	NOUN
bracis-33593	54	6	the	the	DET
bracis-33593	54	7	prediction	prediction	NOUN
bracis-33593	54	8	error	error	NOUN
bracis-33593	54	9	of	of	ADP
bracis-33593	54	10	models	model	NOUN
bracis-33593	54	11	already	already	ADV
bracis-33593	54	12	learned	learn	VERB
bracis-33593	54	13	.	.	PUNCT
bracis-33593	55	1	this	this	DET
bracis-33593	55	2	approach	approach	NOUN
bracis-33593	55	3	resulted	result	VERB
bracis-33593	55	4	in	in	ADP
bracis-33593	55	5	solutions	solution	NOUN
bracis-33593	55	6	that	that	PRON
bracis-33593	55	7	proved	prove	VERB
bracis-33593	55	8	to	to	PART
bracis-33593	55	9	be	be	AUX
bracis-33593	55	10	very	very	ADV
bracis-33593	55	11	suitable	suitable	ADJ
bracis-33593	55	12	for	for	ADP
bracis-33593	55	13	dealing	deal	VERB
bracis-33593	55	14	with	with	ADP
bracis-33593	55	15	non	non	ADJ
bracis-33593	55	16	-	-	ADJ
bracis-33593	55	17	stationary	stationary	ADJ
bracis-33593	55	18	environments	environment	NOUN
bracis-33593	55	19	,	,	PUNCT
bracis-33593	55	20	as	as	SCONJ
bracis-33593	55	21	detailed	detail	VERB
bracis-33593	55	22	by	by	ADP
bracis-33593	55	23	alegre	alegre	PROPN
bracis-33593	55	24	,	,	PUNCT
bracis-33593	55	25	bazzan	bazzan	NOUN
bracis-33593	55	26	e	e	PROPN
bracis-33593	55	27	silva	silva	NOUN
bracis-33593	55	28	(	(	PUNCT
bracis-33593	55	29	2021	2021	NUM
bracis-33593	55	30	)	)	PUNCT
bracis-33593	56	1	[	[	X
bracis-33593	56	2	1	1	X
bracis-33593	56	3	]	]	PUNCT
bracis-33593	56	4	who	who	PRON
bracis-33593	56	5	explored	explore	VERB
bracis-33593	56	6	the	the	DET
bracis-33593	56	7	power	power	NOUN
bracis-33593	56	8	of	of	ADP
bracis-33593	56	9	rapid	rapid	ADJ
bracis-33593	56	10	adaptation	adaptation	NOUN
bracis-33593	56	11	by	by	ADP
bracis-33593	56	12	identifying	identify	VERB
bracis-33593	56	13	change	change	NOUN
bracis-33593	56	14	points	point	NOUN
bracis-33593	56	15	with	with	ADP
bracis-33593	56	16	high	high	ADJ
bracis-33593	56	17	confidence	confidence	NOUN
bracis-33593	56	18	.	.	PUNCT
bracis-33593	57	1	li	li	PROPN
bracis-33593	57	2	et	et	PROPN
bracis-33593	57	3	al	al	PROPN
bracis-33593	57	4	.	.	PUNCT
bracis-33593	58	1	[	[	X
bracis-33593	58	2	7	7	NUM
bracis-33593	58	3	,	,	PUNCT
bracis-33593	58	4	8	8	NUM
bracis-33593	58	5	]	]	PUNCT
bracis-33593	58	6	,	,	PUNCT
bracis-33593	58	7	have	have	AUX
bracis-33593	58	8	delved	delve	VERB
bracis-33593	58	9	into	into	ADP
bracis-33593	58	10	hierarchical	hierarchical	ADJ
bracis-33593	58	11	approaches	approach	NOUN
bracis-33593	58	12	,	,	PUNCT
bracis-33593	58	13	where	where	SCONJ
bracis-33593	58	14	segments	segment	NOUN
bracis-33593	58	15	of	of	ADP
bracis-33593	58	16	information	information	NOUN
bracis-33593	58	17	are	be	AUX
bracis-33593	58	18	compounded	compound	VERB
bracis-33593	58	19	in	in	ADP
bracis-33593	58	20	a	a	DET
bracis-33593	58	21	hierarchical	hierarchical	ADJ
bracis-33593	58	22	manner	manner	NOUN
bracis-33593	58	23	.	.	PUNCT
bracis-33593	59	1	this	this	DET
bracis-33593	59	2	hierarchical	hierarchical	ADJ
bracis-33593	59	3	structuring	structuring	NOUN
bracis-33593	59	4	enables	enable	VERB
bracis-33593	59	5	more	more	ADV
bracis-33593	59	6	efficient	efficient	ADJ
bracis-33593	59	7	learning	learning	NOUN
bracis-33593	59	8	and	and	CCONJ
bracis-33593	59	9	decision	decision	NOUN
bracis-33593	59	10	-	-	PUNCT
bracis-33593	59	11	making	make	VERB
bracis-33593	59	12	processes	process	NOUN
bracis-33593	59	13	,	,	PUNCT
bracis-33593	59	14	particularly	particularly	ADV
bracis-33593	59	15	in	in	ADP
bracis-33593	59	16	complex	complex	ADJ
bracis-33593	59	17	environments	environment	NOUN
bracis-33593	59	18	with	with	ADP
bracis-33593	59	19	multiple	multiple	ADJ
bracis-33593	59	20	levels	level	NOUN
bracis-33593	59	21	of	of	ADP
bracis-33593	59	22	abstraction	abstraction	NOUN
bracis-33593	59	23	.	.	PUNCT
bracis-33593	60	1	there	there	PRON
bracis-33593	60	2	are	be	VERB
bracis-33593	60	3	works	work	NOUN
bracis-33593	60	4	[	[	X
bracis-33593	60	5	10	10	NUM
bracis-33593	60	6	]	]	PUNCT
bracis-33593	60	7	that	that	PRON
bracis-33593	60	8	have	have	AUX
bracis-33593	60	9	focused	focus	VERB
bracis-33593	60	10	on	on	ADP
bracis-33593	60	11	incremental	incremental	ADJ
bracis-33593	60	12	learning	learning	NOUN
bracis-33593	60	13	techniques	technique	NOUN
bracis-33593	60	14	,	,	PUNCT
bracis-33593	60	15	which	which	PRON
bracis-33593	60	16	have	have	AUX
bracis-33593	60	17	shown	show	VERB
bracis-33593	60	18	promising	promising	ADJ
bracis-33593	60	19	results	result	NOUN
bracis-33593	60	20	in	in	ADP
bracis-33593	60	21	adapting	adapt	VERB
bracis-33593	60	22	to	to	ADP
bracis-33593	60	23	changing	change	VERB
bracis-33593	60	24	environments	environment	NOUN
bracis-33593	60	25	and	and	CCONJ
bracis-33593	60	26	evolving	evolving	NOUN
bracis-33593	60	27	tasks	task	NOUN
bracis-33593	60	28	.	.	PUNCT
bracis-33593	61	1	by	by	ADP
bracis-33593	61	2	incrementally	incrementally	ADV
bracis-33593	61	3	updating	update	VERB
bracis-33593	61	4	models	model	NOUN
bracis-33593	61	5	based	base	VERB
bracis-33593	61	6	on	on	ADP
bracis-33593	61	7	new	new	ADJ
bracis-33593	61	8	data	datum	NOUN
bracis-33593	61	9	and	and	CCONJ
bracis-33593	61	10	experiences	experience	NOUN
bracis-33593	61	11	,	,	PUNCT
bracis-33593	61	12	these	these	DET
bracis-33593	61	13	approaches	approach	NOUN
bracis-33593	61	14	offer	offer	VERB
bracis-33593	61	15	flexibility	flexibility	NOUN
bracis-33593	61	16	and	and	CCONJ
bracis-33593	61	17	scalability	scalability	NOUN
bracis-33593	61	18	in	in	ADP
bracis-33593	61	19	learning	learn	VERB
bracis-33593	61	20	complex	complex	ADJ
bracis-33593	61	21	tasks	task	NOUN
bracis-33593	61	22	over	over	ADP
bracis-33593	61	23	time	time	NOUN
bracis-33593	61	24	.	.	PUNCT
bracis-33593	62	1	offline	offline	ADJ
bracis-33593	62	2	learning	learning	NOUN
bracis-33593	62	3	settings	setting	NOUN
bracis-33593	62	4	,	,	PUNCT
bracis-33593	62	5	[	[	X
bracis-33593	62	6	6	6	NUM
bracis-33593	62	7	]	]	PUNCT
bracis-33593	62	8	have	have	AUX
bracis-33593	62	9	demonstrated	demonstrate	VERB
bracis-33593	62	10	the	the	DET
bracis-33593	62	11	effectiveness	effectiveness	NOUN
bracis-33593	62	12	of	of	ADP
bracis-33593	62	13	model	model	NOUN
bracis-33593	62	14	-	-	PUNCT
bracis-33593	62	15	based	base	VERB
bracis-33593	62	16	approaches	approach	NOUN
bracis-33593	62	17	,	,	PUNCT
bracis-33593	62	18	showcasing	showcase	VERB
bracis-33593	62	19	their	their	PRON
bracis-33593	62	20	utility	utility	NOUN
bracis-33593	62	21	in	in	ADP
bracis-33593	62	22	offline	offline	ADJ
bracis-33593	62	23	learning	learn	VERB
bracis-33593	62	24	scenarios	scenario	NOUN
bracis-33593	62	25	where	where	SCONJ
bracis-33593	62	26	access	access	NOUN
bracis-33593	62	27	to	to	ADP
bracis-33593	62	28	real	real	ADJ
bracis-33593	62	29	-	-	PUNCT
bracis-33593	62	30	time	time	NOUN
bracis-33593	62	31	data	datum	NOUN
bracis-33593	62	32	may	may	AUX
bracis-33593	62	33	be	be	AUX
bracis-33593	62	34	limited	limit	VERB
bracis-33593	62	35	or	or	CCONJ
bracis-33593	62	36	impractical	impractical	ADJ
bracis-33593	62	37	.	.	PUNCT
bracis-33593	63	1	leveraging	leverage	VERB
bracis-33593	63	2	offline	offline	ADJ
bracis-33593	63	3	data	datum	NOUN
bracis-33593	63	4	,	,	PUNCT
bracis-33593	63	5	these	these	DET
bracis-33593	63	6	approaches	approach	NOUN
bracis-33593	63	7	enable	enable	VERB
bracis-33593	63	8	efficient	efficient	ADJ
bracis-33593	63	9	learning	learning	NOUN
bracis-33593	63	10	and	and	CCONJ
bracis-33593	63	11	decision	decision	NOUN
bracis-33593	63	12	-	-	PUNCT
bracis-33593	63	13	making	making	NOUN
bracis-33593	63	14	based	base	VERB
bracis-33593	63	15	on	on	ADP
bracis-33593	63	16	previously	previously	ADV
bracis-33593	63	17	collected	collect	VERB
bracis-33593	63	18	experiences	experience	NOUN
bracis-33593	63	19	.	.	PUNCT
bracis-33593	64	1	the	the	DET
bracis-33593	64	2	utilization	utilization	NOUN
bracis-33593	64	3	of	of	ADP
bracis-33593	64	4	a	a	DET
bracis-33593	64	5	likelihood	likelihood	NOUN
bracis-33593	64	6	estimator	estimator	NOUN
bracis-33593	64	7	in	in	ADP
bracis-33593	64	8	the	the	DET
bracis-33593	64	9	context	context	NOUN
bracis-33593	64	10	of	of	ADP
bracis-33593	64	11	building	building	NOUN
bracis-33593	64	12	options	option	NOUN
bracis-33593	64	13	[	[	X
bracis-33593	64	14	4	4	NUM
bracis-33593	64	15	]	]	PUNCT
bracis-33593	64	16	,	,	PUNCT
bracis-33593	64	17	inspired	inspire	VERB
bracis-33593	64	18	our	our	PRON
bracis-33593	64	19	exploration	exploration	NOUN
bracis-33593	64	20	into	into	ADP
bracis-33593	64	21	employing	employ	VERB
bracis-33593	64	22	a	a	DET
bracis-33593	64	23	similar	similar	ADJ
bracis-33593	64	24	approach	approach	NOUN
bracis-33593	64	25	for	for	ADP
bracis-33593	64	26	constructing	construct	VERB
bracis-33593	64	27	models	model	NOUN
bracis-33593	64	28	tailored	tailor	VERB
bracis-33593	64	29	to	to	ADP
bracis-33593	64	30	different	different	ADJ
bracis-33593	64	31	regions	region	NOUN
bracis-33593	64	32	within	within	ADP
bracis-33593	64	33	the	the	DET
bracis-33593	64	34	world	world	NOUN
bracis-33593	64	35	.	.	PUNCT
bracis-33593	65	1	the	the	DET
bracis-33593	65	2	likelihood	likelihood	NOUN
bracis-33593	65	3	estimator	estimator	NOUN
bracis-33593	65	4	framework	framework	NOUN
bracis-33593	65	5	proved	prove	VERB
bracis-33593	65	6	instrumental	instrumental	ADJ
bracis-33593	65	7	in	in	ADP
bracis-33593	65	8	our	our	PRON
bracis-33593	65	9	efforts	effort	NOUN
bracis-33593	65	10	to	to	PART
bracis-33593	65	11	construct	construct	VERB
bracis-33593	65	12	region	region	NOUN
bracis-33593	65	13	-	-	PUNCT
bracis-33593	65	14	specific	specific	ADJ
bracis-33593	65	15	models	model	NOUN
bracis-33593	65	16	,	,	PUNCT
bracis-33593	65	17	enhancing	enhance	VERB
bracis-33593	65	18	the	the	DET
bracis-33593	65	19	efficiency	efficiency	NOUN
bracis-33593	65	20	and	and	CCONJ
bracis-33593	65	21	effectiveness	effectiveness	NOUN
bracis-33593	65	22	of	of	ADP
bracis-33593	65	23	our	our	PRON
bracis-33593	65	24	reinforcement	reinforcement	NOUN
bracis-33593	65	25	learning	learning	NOUN
bracis-33593	65	26	approach	approach	NOUN
bracis-33593	65	27	in	in	ADP
bracis-33593	65	28	complex	complex	ADJ
bracis-33593	65	29	and	and	CCONJ
bracis-33593	65	30	heterogeneous	heterogeneous	ADJ
bracis-33593	65	31	environments	environment	NOUN
bracis-33593	65	32	as	as	SCONJ
bracis-33593	65	33	it	it	PRON
bracis-33593	65	34	was	be	AUX
bracis-33593	65	35	used	use	VERB
bracis-33593	65	36	in	in	ADP
bracis-33593	65	37	the	the	DET
bracis-33593	65	38	options	option	NOUN
bracis-33593	65	39	context	context	NOUN
bracis-33593	65	40	.	.	PUNCT
bracis-33593	66	1	3	3	NUM
bracis-33593	66	2	k	k	NOUN
bracis-33593	66	3	-	-	PUNCT
bracis-33593	66	4	models	model	NOUN
bracis-33593	66	5	mdp	mdp	VERB
bracis-33593	66	6	the	the	DET
bracis-33593	66	7	markov	markov	NOUN
bracis-33593	66	8	decision	decision	NOUN
bracis-33593	66	9	process	process	NOUN
bracis-33593	66	10	(	(	PUNCT
bracis-33593	66	11	mdp	mdp	NOUN
bracis-33593	66	12	)	)	PUNCT
bracis-33593	67	1	[	[	X
bracis-33593	67	2	13	13	NUM
bracis-33593	67	3	]	]	PUNCT
bracis-33593	67	4	framework	framework	NOUN
bracis-33593	67	5	serves	serve	VERB
bracis-33593	67	6	as	as	ADP
bracis-33593	67	7	the	the	DET
bracis-33593	67	8	foundational	foundational	ADJ
bracis-33593	67	9	formalism	formalism	NOUN
bracis-33593	67	10	for	for	ADP
bracis-33593	67	11	representing	represent	VERB
bracis-33593	67	12	the	the	DET
bracis-33593	67	13	core	core	NOUN
bracis-33593	67	14	aspects	aspect	NOUN
bracis-33593	67	15	of	of	ADP
bracis-33593	67	16	reinforcement	reinforcement	NOUN
bracis-33593	67	17	learning	learning	NOUN
bracis-33593	67	18	.	.	PUNCT
bracis-33593	68	1	an	an	DET
bracis-33593	68	2	mdp	mdp	NOUN
bracis-33593	68	3	is	be	AUX
bracis-33593	68	4	typically	typically	ADV
bracis-33593	68	5	characterized	characterize	VERB
bracis-33593	68	6	by	by	ADP
bracis-33593	68	7	the	the	DET
bracis-33593	68	8	tuple	tuple	PROPN
bracis-33593	68	9	\(<\mathcal	\(<\mathcal	ADJ
bracis-33593	68	10	{	{	PUNCT
bracis-33593	68	11	s	s	PROPN
bracis-33593	68	12	}	}	PUNCT
bracis-33593	68	13	,	,	PUNCT
bracis-33593	68	14	\mathcal	\mathcal	PROPN
bracis-33593	68	15	{	{	PUNCT
bracis-33593	68	16	a	a	NOUN
bracis-33593	68	17	}	}	PUNCT
bracis-33593	68	18	,	,	PUNCT
bracis-33593	68	19	\mathcal	\mathcal	PROPN
bracis-33593	68	20	{	{	PUNCT
bracis-33593	68	21	p	p	NOUN
bracis-33593	68	22	}	}	PUNCT
bracis-33593	68	23	,	,	PUNCT
bracis-33593	68	24	\mathcal	\mathcal	PROPN
bracis-33593	68	25	{	{	PUNCT
bracis-33593	68	26	r}>\	r}>\	PROPN
bracis-33593	68	27	)	)	PUNCT
bracis-33593	68	28	,	,	PUNCT
bracis-33593	68	29	where	where	SCONJ
bracis-33593	68	30	\(\mathcal	\(\mathcal	ADJ
bracis-33593	68	31	{	{	PUNCT
bracis-33593	68	32	s}\	s}\	PROPN
bracis-33593	68	33	)	)	PUNCT
bracis-33593	68	34	denotes	denote	VERB
bracis-33593	68	35	the	the	DET
bracis-33593	68	36	set	set	NOUN
bracis-33593	68	37	of	of	ADP
bracis-33593	68	38	states	state	NOUN
bracis-33593	68	39	,	,	PUNCT
bracis-33593	68	40	\(\mathcal	\(\mathcal	ADJ
bracis-33593	68	41	{	{	PUNCT
bracis-33593	68	42	a}\	a}\	NUM
bracis-33593	68	43	)	)	PUNCT
bracis-33593	68	44	represents	represent	VERB
bracis-33593	68	45	the	the	DET
bracis-33593	68	46	set	set	NOUN
bracis-33593	68	47	of	of	ADP
bracis-33593	68	48	actions	action	NOUN
bracis-33593	68	49	,	,	PUNCT
bracis-33593	68	50	\(\mathcal	\(\mathcal	ADJ
bracis-33593	68	51	{	{	PUNCT
bracis-33593	68	52	p}\	p}\	NOUN
bracis-33593	68	53	)	)	PUNCT
bracis-33593	68	54	defines	define	VERB
bracis-33593	68	55	the	the	DET
bracis-33593	68	56	transition	transition	NOUN
bracis-33593	68	57	probabilities	probability	NOUN
bracis-33593	68	58	between	between	ADP
bracis-33593	68	59	states	state	NOUN
bracis-33593	68	60	given	give	VERB
bracis-33593	68	61	actions	action	NOUN
bracis-33593	68	62	,	,	PUNCT
bracis-33593	68	63	and	and	CCONJ
bracis-33593	68	64	\(\mathcal	\(\mathcal	ADJ
bracis-33593	68	65	{	{	PUNCT
bracis-33593	68	66	r}\	r}\	NOUN
bracis-33593	68	67	)	)	PUNCT
bracis-33593	68	68	denotes	denote	VERB
bracis-33593	68	69	the	the	DET
bracis-33593	68	70	reward	reward	NOUN
bracis-33593	68	71	function	function	NOUN
bracis-33593	68	72	associated	associate	VERB
bracis-33593	68	73	with	with	ADP
bracis-33593	68	74	each	each	DET
bracis-33593	68	75	state	state	NOUN
bracis-33593	68	76	.	.	PUNCT
bracis-33593	69	1	if	if	SCONJ
bracis-33593	69	2	we	we	PRON
bracis-33593	69	3	assume	assume	VERB
bracis-33593	69	4	a	a	DET
bracis-33593	69	5	know	know	NOUN
bracis-33593	69	6	\(\mathcal	\(\mathcal	ADJ
bracis-33593	69	7	{	{	PUNCT
bracis-33593	69	8	p}\	p}\	NOUN
bracis-33593	69	9	)	)	PUNCT
bracis-33593	69	10	parameterized	parameterized	ADJ
bracis-33593	69	11	function	function	VERB
bracis-33593	69	12	a	a	DET
bracis-33593	69	13	natural	natural	ADJ
bracis-33593	69	14	implementation	implementation	NOUN
bracis-33593	69	15	for	for	ADP
bracis-33593	69	16	a	a	DET
bracis-33593	69	17	model	model	NOUN
bracis-33593	69	18	based	base	VERB
bracis-33593	69	19	approach	approach	NOUN
bracis-33593	69	20	would	would	AUX
bracis-33593	69	21	be	be	AUX
bracis-33593	69	22	to	to	PART
bracis-33593	69	23	find	find	VERB
bracis-33593	69	24	the	the	DET
bracis-33593	69	25	parameters	parameter	NOUN
bracis-33593	69	26	that	that	PRON
bracis-33593	69	27	fit	fit	VERB
bracis-33593	69	28	a	a	DET
bracis-33593	69	29	observed	observed	ADJ
bracis-33593	69	30	experience	experience	NOUN
bracis-33593	69	31	,	,	PUNCT
bracis-33593	69	32	and	and	CCONJ
bracis-33593	69	33	if	if	SCONJ
bracis-33593	69	34	we	we	PRON
bracis-33593	69	35	assume	assume	VERB
bracis-33593	69	36	that	that	SCONJ
bracis-33593	69	37	the	the	DET
bracis-33593	69	38	dynamics	dynamic	NOUN
bracis-33593	69	39	of	of	ADP
bracis-33593	69	40	the	the	DET
bracis-33593	69	41	world	world	NOUN
bracis-33593	69	42	vary	vary	VERB
bracis-33593	69	43	spatially	spatially	ADV
bracis-33593	69	44	,	,	PUNCT
bracis-33593	69	45	meaning	mean	VERB
bracis-33593	69	46	that	that	SCONJ
bracis-33593	69	47	different	different	ADJ
bracis-33593	69	48	regions	region	NOUN
bracis-33593	69	49	exhibit	exhibit	VERB
bracis-33593	69	50	distinct	distinct	ADJ
bracis-33593	69	51	behaviors	behavior	NOUN
bracis-33593	69	52	,	,	PUNCT
bracis-33593	69	53	we	we	PRON
bracis-33593	69	54	can	can	AUX
bracis-33593	69	55	define	define	VERB
bracis-33593	69	56	a	a	DET
bracis-33593	69	57	function	function	NOUN
bracis-33593	69	58	\(\beta	\(\beta	NOUN
bracis-33593	69	59	:	:	PUNCT
bracis-33593	69	60	\mathcal	\mathcal	PROPN
bracis-33593	69	61	{	{	PUNCT
bracis-33593	69	62	s	s	PROPN
bracis-33593	69	63	}	}	PUNCT
bracis-33593	69	64	\mapsto	\mapsto	PROPN
bracis-33593	69	65	\theta	\theta	PROPN
bracis-33593	69	66	\	\	PROPN
bracis-33593	69	67	)	)	PUNCT
bracis-33593	69	68	that	that	PRON
bracis-33593	69	69	give	give	VERB
bracis-33593	69	70	us	we	PRON
bracis-33593	69	71	the	the	DET
bracis-33593	69	72	parameter	parameter	NOUN
bracis-33593	69	73	for	for	ADP
bracis-33593	69	74	the	the	DET
bracis-33593	69	75	model	model	NOUN
bracis-33593	69	76	based	base	VERB
bracis-33593	69	77	on	on	ADP
bracis-33593	69	78	a	a	DET
bracis-33593	69	79	given	give	VERB
bracis-33593	69	80	state	state	NOUN
bracis-33593	69	81	.	.	PUNCT
bracis-33593	70	1	we	we	PRON
bracis-33593	70	2	are	be	AUX
bracis-33593	70	3	going	go	VERB
bracis-33593	70	4	to	to	PART
bracis-33593	70	5	extend	extend	VERB
bracis-33593	70	6	this	this	DET
bracis-33593	70	7	concept	concept	NOUN
bracis-33593	70	8	now	now	ADV
bracis-33593	70	9	based	base	VERB
bracis-33593	70	10	in	in	ADP
bracis-33593	70	11	a	a	DET
bracis-33593	70	12	assumption	assumption	NOUN
bracis-33593	70	13	that	that	SCONJ
bracis-33593	70	14	the	the	DET
bracis-33593	70	15	world	world	NOUN
bracis-33593	70	16	has	have	AUX
bracis-33593	70	17	k	k	ADJ
bracis-33593	70	18	regions	region	NOUN
bracis-33593	70	19	where	where	SCONJ
bracis-33593	70	20	the	the	DET
bracis-33593	70	21	parameters	parameter	NOUN
bracis-33593	70	22	can	can	AUX
bracis-33593	70	23	be	be	AUX
bracis-33593	70	24	different	different	ADJ
bracis-33593	70	25	in	in	ADP
bracis-33593	70	26	the	the	DET
bracis-33593	70	27	context	context	NOUN
bracis-33593	70	28	of	of	ADP
bracis-33593	70	29	modeling	model	VERB
bracis-33593	70	30	segmented	segment	VERB
bracis-33593	70	31	dynamics	dynamic	NOUN
bracis-33593	70	32	of	of	ADP
bracis-33593	70	33	the	the	DET
bracis-33593	70	34	world	world	NOUN
bracis-33593	70	35	,	,	PUNCT
bracis-33593	70	36	we	we	PRON
bracis-33593	70	37	extend	extend	VERB
bracis-33593	70	38	the	the	DET
bracis-33593	70	39	mdp	mdp	NOUN
bracis-33593	70	40	framework	framework	NOUN
bracis-33593	70	41	to	to	ADP
bracis-33593	70	42	\(<\mathcal	\(<\mathcal	ADJ
bracis-33593	70	43	{	{	PUNCT
bracis-33593	70	44	s	s	X
bracis-33593	70	45	}	}	PUNCT
bracis-33593	70	46	,	,	PUNCT
bracis-33593	70	47	\mathcal	\mathcal	PROPN
bracis-33593	70	48	{	{	PUNCT
bracis-33593	70	49	a	a	NOUN
bracis-33593	70	50	}	}	PUNCT
bracis-33593	70	51	,	,	PUNCT
bracis-33593	70	52	\mathcal	\mathcal	PROPN
bracis-33593	70	53	{	{	PUNCT
bracis-33593	70	54	p	p	NOUN
bracis-33593	70	55	}	}	PUNCT
bracis-33593	70	56	,	,	PUNCT
bracis-33593	70	57	\mathcal	\mathcal	X
bracis-33593	70	58	{	{	PUNCT
bracis-33593	70	59	r	r	NOUN
bracis-33593	70	60	}	}	PUNCT
bracis-33593	70	61	,	,	PUNCT
bracis-33593	70	62	\vartheta	\vartheta	INTJ
bracis-33593	70	63	,	,	PUNCT
bracis-33593	70	64	\beta	\beta	PROPN
bracis-33593	70	65	,	,	PUNCT
bracis-33593	70	66	\mathcal	\mathcal	ADJ
bracis-33593	70	67	{	{	PUNCT
bracis-33593	70	68	g	g	NOUN
bracis-33593	70	69	}	}	PUNCT
bracis-33593	70	70	,	,	PUNCT
bracis-33593	70	71	\mathcal	\mathcal	PROPN
bracis-33593	70	72	{	{	PUNCT
bracis-33593	70	73	s}_0>\	s}_0>\	NOUN
bracis-33593	70	74	)	)	PUNCT
bracis-33593	70	75	where	where	SCONJ
bracis-33593	70	76	the	the	DET
bracis-33593	70	77	transition	transition	NOUN
bracis-33593	70	78	function	function	VERB
bracis-33593	70	79	\(\mathcal	\(\mathcal	ADJ
bracis-33593	70	80	{	{	PUNCT
bracis-33593	70	81	p}\	p}\	NOUN
bracis-33593	70	82	)	)	PUNCT
bracis-33593	70	83	is	be	AUX
bracis-33593	70	84	parameterized	parameterize	VERB
bracis-33593	70	85	by	by	ADP
bracis-33593	70	86	\(\theta	\(\theta	NOUN
bracis-33593	70	87	\	\	PROPN
bracis-33593	70	88	)	)	PUNCT
bracis-33593	70	89	,	,	PUNCT
bracis-33593	70	90	so	so	CCONJ
bracis-33593	70	91	a	a	DET
bracis-33593	70	92	segment	segment	NOUN
bracis-33593	70	93	of	of	ADP
bracis-33593	70	94	the	the	DET
bracis-33593	70	95	world	world	NOUN
bracis-33593	70	96	is	be	AUX
bracis-33593	70	97	represented	represent	VERB
bracis-33593	70	98	by	by	ADP
bracis-33593	70	99	a	a	DET
bracis-33593	70	100	\(\theta	\(\theta	NOUN
bracis-33593	70	101	\	\	NOUN
bracis-33593	70	102	)	)	PUNCT
bracis-33593	70	103	that	that	PRON
bracis-33593	70	104	dictates	dictate	VERB
bracis-33593	70	105	the	the	DET
bracis-33593	70	106	dynamics	dynamic	NOUN
bracis-33593	70	107	applied	apply	VERB
bracis-33593	70	108	by	by	ADP
bracis-33593	70	109	\(\mathcal	\(\mathcal	ADJ
bracis-33593	70	110	{	{	PUNCT
bracis-33593	70	111	p}\	p}\	NOUN
bracis-33593	70	112	)	)	PUNCT
bracis-33593	70	113	on	on	ADP
bracis-33593	70	114	that	that	DET
bracis-33593	70	115	region	region	NOUN
bracis-33593	70	116	of	of	ADP
bracis-33593	70	117	the	the	DET
bracis-33593	70	118	state	state	NOUN
bracis-33593	70	119	space	space	NOUN
bracis-33593	70	120	,	,	PUNCT
bracis-33593	70	121	and	and	CCONJ
bracis-33593	70	122	to	to	PART
bracis-33593	70	123	define	define	VERB
bracis-33593	70	124	that	that	DET
bracis-33593	70	125	region	region	NOUN
bracis-33593	70	126	we	we	PRON
bracis-33593	70	127	introduce	introduce	VERB
bracis-33593	70	128	a	a	DET
bracis-33593	70	129	new	new	ADJ
bracis-33593	70	130	function	function	NOUN
bracis-33593	70	131	\(\beta	\(\beta	NOUN
bracis-33593	70	132	:	:	PUNCT
bracis-33593	70	133	\mathcal	\mathcal	PROPN
bracis-33593	70	134	{	{	PUNCT
bracis-33593	70	135	s	s	NOUN
bracis-33593	70	136	}	}	PUNCT
bracis-33593	70	137	\times	\times	ADP
bracis-33593	70	138	\theta	\theta	PROPN
bracis-33593	70	139	\mapsto	\mapsto	PROPN
bracis-33593	71	1	[	[	X
bracis-33593	71	2	0,1]\	0,1]\	NOUN
bracis-33593	71	3	)	)	PUNCT
bracis-33593	71	4	,	,	PUNCT
bracis-33593	71	5	which	which	PRON
bracis-33593	71	6	assigns	assign	VERB
bracis-33593	71	7	a	a	DET
bracis-33593	71	8	probability	probability	NOUN
bracis-33593	71	9	of	of	ADP
bracis-33593	71	10	the	the	DET
bracis-33593	71	11	parameters	parameter	NOUN
bracis-33593	71	12	\(\theta	\(\theta	NOUN
bracis-33593	71	13	\	\	PROPN
bracis-33593	71	14	)	)	PUNCT
bracis-33593	71	15	based	base	VERB
bracis-33593	71	16	on	on	ADP
bracis-33593	71	17	the	the	DET
bracis-33593	71	18	current	current	ADJ
bracis-33593	71	19	state	state	NOUN
bracis-33593	71	20	.	.	PUNCT
bracis-33593	72	1	so	so	ADV
bracis-33593	72	2	given	give	VERB
bracis-33593	72	3	a	a	DET
bracis-33593	72	4	state	state	NOUN
bracis-33593	72	5	s	s	NOUN
bracis-33593	72	6	and	and	CCONJ
bracis-33593	72	7	a	a	DET
bracis-33593	72	8	parameter	parameter	NOUN
bracis-33593	72	9	\(\theta	\(\theta	NOUN
bracis-33593	72	10	\	\	PROPN
bracis-33593	72	11	)	)	PUNCT
bracis-33593	72	12	the	the	DET
bracis-33593	72	13	\(\beta	\(\beta	PROPN
bracis-33593	72	14	(	(	PUNCT
bracis-33593	72	15	s	s	PROPN
bracis-33593	72	16	,	,	PUNCT
bracis-33593	72	17	\theta	\theta	PROPN
bracis-33593	72	18	)	)	PUNCT
bracis-33593	72	19	\	\	NOUN
bracis-33593	72	20	)	)	PUNCT
bracis-33593	72	21	function	function	NOUN
bracis-33593	72	22	yields	yield	VERB
bracis-33593	72	23	the	the	DET
bracis-33593	72	24	probabilities	probability	NOUN
bracis-33593	72	25	for	for	ADP
bracis-33593	72	26	parameter	parameter	NOUN
bracis-33593	72	27	\(\theta	\(\theta	NOUN
bracis-33593	72	28	\	\	PROPN
bracis-33593	72	29	)	)	PUNCT
bracis-33593	72	30	to	to	PART
bracis-33593	72	31	be	be	AUX
bracis-33593	72	32	the	the	DET
bracis-33593	72	33	one	one	NOUN
bracis-33593	72	34	that	that	PRON
bracis-33593	72	35	parameterize	parameterize	VERB
bracis-33593	72	36	\(\mathcal	\(\mathcal	ADJ
bracis-33593	72	37	{	{	PUNCT
bracis-33593	72	38	p}\	p}\	NOUN
bracis-33593	72	39	)	)	PUNCT
bracis-33593	72	40	on	on	ADP
bracis-33593	72	41	s.	s.	PROPN
bracis-33593	72	42	the	the	DET
bracis-33593	72	43	set	set	PROPN
bracis-33593	72	44	\(\vartheta	\(\vartheta	NOUN
bracis-33593	72	45	\	\	NOUN
bracis-33593	72	46	)	)	PUNCT
bracis-33593	72	47	represents	represent	VERB
bracis-33593	72	48	the	the	DET
bracis-33593	72	49	union	union	NOUN
bracis-33593	72	50	of	of	ADP
bracis-33593	72	51	all	all	DET
bracis-33593	72	52	possible	possible	ADJ
bracis-33593	72	53	parameters	parameter	NOUN
bracis-33593	72	54	evolved	evolve	VERB
bracis-33593	72	55	in	in	ADP
bracis-33593	72	56	functions	function	NOUN
bracis-33593	72	57	\(\mathcal	\(\mathcal	ADJ
bracis-33593	72	58	{	{	PUNCT
bracis-33593	72	59	p}\	p}\	NOUN
bracis-33593	72	60	)	)	PUNCT
bracis-33593	72	61	that	that	PRON
bracis-33593	72	62	maps	map	VERB
bracis-33593	72	63	a	a	DET
bracis-33593	72	64	region	region	NOUN
bracis-33593	72	65	in	in	ADP
bracis-33593	72	66	the	the	DET
bracis-33593	72	67	environment	environment	NOUN
bracis-33593	72	68	,	,	PUNCT
bracis-33593	72	69	in	in	ADP
bracis-33593	72	70	a	a	DET
bracis-33593	72	71	way	way	NOUN
bracis-33593	72	72	that	that	PRON
bracis-33593	72	73	for	for	ADP
bracis-33593	72	74	a	a	DET
bracis-33593	72	75	world	world	NOUN
bracis-33593	72	76	with	with	ADP
bracis-33593	72	77	k	k	PROPN
bracis-33593	72	78	distinct	distinct	ADJ
bracis-33593	72	79	regions	region	NOUN
bracis-33593	72	80	\(\vartheta	\(\vartheta	NOUN
bracis-33593	72	81	=	=	PUNCT
bracis-33593	72	82	\{\theta	\{\theta	NUM
bracis-33593	72	83	_	_	NOUN
bracis-33593	72	84	1	1	NUM
bracis-33593	72	85	,	,	PUNCT
bracis-33593	72	86	\theta	\theta	PROPN
bracis-33593	72	87	_	_	PROPN
bracis-33593	72	88	2	2	NUM
bracis-33593	72	89	...	...	PUNCT
bracis-33593	72	90	,	,	PUNCT
bracis-33593	72	91	\theta	\theta	PROPN
bracis-33593	72	92	_	_	PROPN
bracis-33593	72	93	k\}\	k\}\	NOUN
bracis-33593	72	94	)	)	PUNCT
bracis-33593	72	95	.	.	PUNCT
bracis-33593	73	1	each	each	DET
bracis-33593	73	2	\(\theta	\(\theta	NOUN
bracis-33593	73	3	_	_	PUNCT
bracis-33593	73	4	i\	i\	NOUN
bracis-33593	73	5	)	)	PUNCT
bracis-33593	73	6	represents	represent	VERB
bracis-33593	73	7	a	a	DET
bracis-33593	73	8	distinct	distinct	ADJ
bracis-33593	73	9	parameter	parameter	NOUN
bracis-33593	73	10	for	for	ADP
bracis-33593	73	11	\(\mathcal	\(\mathcal	ADJ
bracis-33593	73	12	{	{	PUNCT
bracis-33593	73	13	p}\	p}\	NOUN
bracis-33593	73	14	)	)	PUNCT
bracis-33593	73	15	that	that	PRON
bracis-33593	73	16	captures	capture	VERB
bracis-33593	73	17	the	the	DET
bracis-33593	73	18	dynamics	dynamic	NOUN
bracis-33593	73	19	of	of	ADP
bracis-33593	73	20	the	the	DET
bracis-33593	73	21	environment	environment	NOUN
bracis-33593	73	22	within	within	ADP
bracis-33593	73	23	a	a	DET
bracis-33593	73	24	specific	specific	ADJ
bracis-33593	73	25	region	region	NOUN
bracis-33593	73	26	.	.	PUNCT
bracis-33593	74	1	we	we	PRON
bracis-33593	74	2	define	define	VERB
bracis-33593	74	3	\(\mathcal	\(\mathcal	ADJ
bracis-33593	74	4	{	{	PUNCT
bracis-33593	74	5	g	g	NOUN
bracis-33593	74	6	}	}	PUNCT
bracis-33593	74	7	\subset	\subset	PROPN
bracis-33593	74	8	\mathcal	\mathcal	PROPN
bracis-33593	74	9	{	{	PUNCT
bracis-33593	74	10	s}\	s}\	PROPN
bracis-33593	74	11	)	)	PUNCT
bracis-33593	74	12	as	as	SCONJ
bracis-33593	74	13	the	the	DET
bracis-33593	74	14	set	set	NOUN
bracis-33593	74	15	of	of	ADP
bracis-33593	74	16	goal	goal	NOUN
bracis-33593	74	17	states	state	NOUN
bracis-33593	74	18	used	use	VERB
bracis-33593	74	19	by	by	ADP
bracis-33593	74	20	the	the	DET
bracis-33593	74	21	planning	planning	NOUN
bracis-33593	74	22	module	module	NOUN
bracis-33593	74	23	to	to	PART
bracis-33593	74	24	construct	construct	VERB
bracis-33593	74	25	plans	plan	NOUN
bracis-33593	74	26	to	to	PART
bracis-33593	74	27	reach	reach	VERB
bracis-33593	74	28	them	they	PRON
bracis-33593	74	29	,	,	PUNCT
bracis-33593	74	30	and	and	CCONJ
bracis-33593	74	31	\(\mathcal	\(\mathcal	ADJ
bracis-33593	74	32	{	{	PUNCT
bracis-33593	74	33	s}_0	s}_0	PROPN
bracis-33593	74	34	\subset	\subset	PROPN
bracis-33593	74	35	\mathcal	\mathcal	PROPN
bracis-33593	74	36	{	{	PUNCT
bracis-33593	74	37	s}\	s}\	PROPN
bracis-33593	74	38	)	)	PUNCT
bracis-33593	74	39	as	as	ADP
bracis-33593	74	40	the	the	DET
bracis-33593	74	41	set	set	NOUN
bracis-33593	74	42	of	of	ADP
bracis-33593	74	43	initial	initial	ADJ
bracis-33593	74	44	states	state	NOUN
bracis-33593	74	45	used	use	VERB
bracis-33593	74	46	to	to	PART
bracis-33593	74	47	randomly	randomly	ADV
bracis-33593	74	48	define	define	VERB
bracis-33593	74	49	the	the	DET
bracis-33593	74	50	initial	initial	ADJ
bracis-33593	74	51	state	state	NOUN
bracis-33593	74	52	of	of	ADP
bracis-33593	74	53	each	each	DET
bracis-33593	74	54	episode	episode	NOUN
bracis-33593	74	55	.	.	PUNCT
bracis-33593	75	1	4	4	NUM
bracis-33593	75	2	likelihood	likelihood	NOUN
bracis-33593	75	3	solution	solution	NOUN
bracis-33593	75	4	for	for	ADP
bracis-33593	75	5	 	 	SPACE
bracis-33593	75	6	a	a	DET
bracis-33593	75	7	 	 	SPACE
bracis-33593	75	8	k	k	NOUN
bracis-33593	75	9	-	-	PUNCT
bracis-33593	75	10	models	model	NOUN
bracis-33593	75	11	mdp	mdp	VERB
bracis-33593	75	12	the	the	DET
bracis-33593	75	13	model	model	NOUN
bracis-33593	75	14	-	-	PUNCT
bracis-33593	75	15	based	base	VERB
bracis-33593	75	16	approach	approach	NOUN
bracis-33593	75	17	to	to	ADP
bracis-33593	75	18	solving	solve	VERB
bracis-33593	75	19	mdps	mdps	NOUN
bracis-33593	75	20	involves	involve	VERB
bracis-33593	75	21	learning	learn	VERB
bracis-33593	75	22	a	a	DET
bracis-33593	75	23	model	model	NOUN
bracis-33593	75	24	that	that	PRON
bracis-33593	75	25	describes	describe	VERB
bracis-33593	75	26	the	the	DET
bracis-33593	75	27	transitions	transition	NOUN
bracis-33593	75	28	between	between	ADP
bracis-33593	75	29	states	state	NOUN
bracis-33593	75	30	in	in	ADP
bracis-33593	75	31	the	the	DET
bracis-33593	75	32	environment	environment	NOUN
bracis-33593	75	33	.	.	PUNCT
bracis-33593	76	1	this	this	DET
bracis-33593	76	2	model	model	NOUN
bracis-33593	76	3	is	be	AUX
bracis-33593	76	4	acquired	acquire	VERB
bracis-33593	76	5	by	by	ADP
bracis-33593	76	6	observing	observe	VERB
bracis-33593	76	7	real	real	ADJ
bracis-33593	76	8	transitions	transition	NOUN
bracis-33593	76	9	that	that	PRON
bracis-33593	76	10	occur	occur	VERB
bracis-33593	76	11	as	as	SCONJ
bracis-33593	76	12	the	the	DET
bracis-33593	76	13	agent	agent	NOUN
bracis-33593	76	14	interacts	interact	VERB
bracis-33593	76	15	with	with	ADP
bracis-33593	76	16	the	the	DET
bracis-33593	76	17	environment	environment	NOUN
bracis-33593	76	18	.	.	PUNCT
bracis-33593	77	1	over	over	ADP
bracis-33593	77	2	time	time	NOUN
bracis-33593	77	3	,	,	PUNCT
bracis-33593	77	4	as	as	SCONJ
bracis-33593	77	5	more	more	ADJ
bracis-33593	77	6	experience	experience	NOUN
bracis-33593	77	7	is	be	AUX
bracis-33593	77	8	gathered	gather	VERB
bracis-33593	77	9	,	,	PUNCT
bracis-33593	77	10	the	the	DET
bracis-33593	77	11	model	model	NOUN
bracis-33593	77	12	can	can	AUX
bracis-33593	77	13	be	be	AUX
bracis-33593	77	14	refined	refine	VERB
bracis-33593	77	15	and	and	CCONJ
bracis-33593	77	16	adjusted	adjust	VERB
bracis-33593	77	17	to	to	ADP
bracis-33593	77	18	more	more	ADV
bracis-33593	77	19	accurately	accurately	ADV
bracis-33593	77	20	reflect	reflect	VERB
bracis-33593	77	21	the	the	DET
bracis-33593	77	22	dynamics	dynamic	NOUN
bracis-33593	77	23	of	of	ADP
bracis-33593	77	24	the	the	DET
bracis-33593	77	25	environment	environment	NOUN
bracis-33593	77	26	.	.	PUNCT
bracis-33593	78	1	the	the	DET
bracis-33593	78	2	process	process	NOUN
bracis-33593	78	3	of	of	ADP
bracis-33593	78	4	learning	learn	VERB
bracis-33593	78	5	the	the	DET
bracis-33593	78	6	model	model	NOUN
bracis-33593	78	7	typically	typically	ADV
bracis-33593	78	8	involves	involve	VERB
bracis-33593	78	9	estimating	estimate	VERB
bracis-33593	78	10	transition	transition	NOUN
bracis-33593	78	11	probabilities	probability	NOUN
bracis-33593	78	12	or	or	CCONJ
bracis-33593	78	13	dynamics	dynamic	NOUN
bracis-33593	78	14	from	from	ADP
bracis-33593	78	15	observed	observed	ADJ
bracis-33593	78	16	state	state	NOUN
bracis-33593	78	17	-	-	PUNCT
bracis-33593	78	18	action	action	NOUN
bracis-33593	78	19	-	-	PUNCT
bracis-33593	78	20	state	state	NOUN
bracis-33593	78	21	transitions	transition	NOUN
bracis-33593	78	22	.	.	PUNCT
bracis-33593	79	1	we	we	PRON
bracis-33593	79	2	bring	bring	VERB
bracis-33593	79	3	a	a	DET
bracis-33593	79	4	likelihood	likelihood	NOUN
bracis-33593	79	5	approach	approach	NOUN
bracis-33593	79	6	to	to	PART
bracis-33593	79	7	find	find	VERB
bracis-33593	79	8	the	the	DET
bracis-33593	79	9	best	good	ADJ
bracis-33593	79	10	parameters	parameter	NOUN
bracis-33593	79	11	of	of	ADP
bracis-33593	79	12	the	the	DET
bracis-33593	79	13	model	model	NOUN
bracis-33593	79	14	based	base	VERB
bracis-33593	79	15	on	on	ADP
bracis-33593	79	16	the	the	DET
bracis-33593	79	17	observed	observed	ADJ
bracis-33593	79	18	experience	experience	NOUN
bracis-33593	79	19	.	.	PUNCT
bracis-33593	80	1	as	as	SCONJ
bracis-33593	80	2	the	the	DET
bracis-33593	80	3	agent	agent	NOUN
bracis-33593	80	4	interacts	interact	VERB
bracis-33593	80	5	with	with	ADP
bracis-33593	80	6	the	the	DET
bracis-33593	80	7	environment	environment	NOUN
bracis-33593	80	8	and	and	CCONJ
bracis-33593	80	9	accumulates	accumulate	VERB
bracis-33593	80	10	more	more	ADJ
bracis-33593	80	11	experience	experience	NOUN
bracis-33593	80	12	,	,	PUNCT
bracis-33593	80	13	we	we	PRON
bracis-33593	80	14	can	can	AUX
bracis-33593	80	15	update	update	VERB
bracis-33593	80	16	the	the	DET
bracis-33593	80	17	parameters	parameter	NOUN
bracis-33593	80	18	so	so	SCONJ
bracis-33593	80	19	the	the	DET
bracis-33593	80	20	learned	learn	VERB
bracis-33593	80	21	model	model	NOUN
bracis-33593	80	22	becomes	become	VERB
bracis-33593	80	23	increasingly	increasingly	ADV
bracis-33593	80	24	accurate	accurate	ADJ
bracis-33593	80	25	and	and	CCONJ
bracis-33593	80	26	reliable	reliable	ADJ
bracis-33593	80	27	.	.	PUNCT
bracis-33593	81	1	by	by	ADP
bracis-33593	81	2	learning	learn	VERB
bracis-33593	81	3	and	and	CCONJ
bracis-33593	81	4	utilizing	utilize	VERB
bracis-33593	81	5	this	this	DET
bracis-33593	81	6	model	model	NOUN
bracis-33593	81	7	,	,	PUNCT
bracis-33593	81	8	the	the	DET
bracis-33593	81	9	agent	agent	NOUN
bracis-33593	81	10	can	can	AUX
bracis-33593	81	11	plan	plan	VERB
bracis-33593	81	12	and	and	CCONJ
bracis-33593	81	13	make	make	VERB
bracis-33593	81	14	decisions	decision	NOUN
bracis-33593	81	15	more	more	ADV
bracis-33593	81	16	effectively	effectively	ADV
bracis-33593	81	17	,	,	PUNCT
bracis-33593	81	18	for	for	ADP
bracis-33593	81	19	solving	solve	VERB
bracis-33593	81	20	the	the	DET
bracis-33593	81	21	reinforcement	reinforcement	NOUN
bracis-33593	81	22	learning	learning	NOUN
bracis-33593	81	23	problem	problem	NOUN
bracis-33593	81	24	at	at	ADP
bracis-33593	81	25	hand	hand	NOUN
bracis-33593	81	26	.	.	PUNCT
bracis-33593	82	1	to	to	PART
bracis-33593	82	2	facilitate	facilitate	VERB
bracis-33593	82	3	performance	performance	NOUN
bracis-33593	82	4	comparison	comparison	NOUN
bracis-33593	82	5	,	,	PUNCT
bracis-33593	82	6	we	we	PRON
bracis-33593	82	7	have	have	AUX
bracis-33593	82	8	implemented	implement	VERB
bracis-33593	82	9	a	a	DET
bracis-33593	82	10	separate	separate	ADJ
bracis-33593	82	11	planner	planner	NOUN
bracis-33593	82	12	module	module	NOUN
bracis-33593	82	13	capable	capable	ADJ
bracis-33593	82	14	of	of	ADP
bracis-33593	82	15	generating	generate	VERB
bracis-33593	82	16	plan	plan	NOUN
bracis-33593	82	17	sequences	sequence	NOUN
bracis-33593	82	18	to	to	PART
bracis-33593	82	19	reach	reach	VERB
bracis-33593	82	20	a	a	DET
bracis-33593	82	21	goal	goal	NOUN
bracis-33593	82	22	state	state	NOUN
bracis-33593	82	23	given	give	VERB
bracis-33593	82	24	a	a	DET
bracis-33593	82	25	model	model	NOUN
bracis-33593	82	26	of	of	ADP
bracis-33593	82	27	the	the	DET
bracis-33593	82	28	world	world	NOUN
bracis-33593	82	29	.	.	PUNCT
bracis-33593	83	1	this	this	DET
bracis-33593	83	2	planner	planner	NOUN
bracis-33593	83	3	module	module	NOUN
bracis-33593	83	4	can	can	AUX
bracis-33593	83	5	be	be	AUX
bracis-33593	83	6	utilized	utilize	VERB
bracis-33593	83	7	with	with	ADP
bracis-33593	83	8	various	various	ADJ
bracis-33593	83	9	implementations	implementation	NOUN
bracis-33593	83	10	,	,	PUNCT
bracis-33593	83	11	including	include	VERB
bracis-33593	83	12	the	the	DET
bracis-33593	83	13	k	k	NOUN
bracis-33593	83	14	-	-	PUNCT
bracis-33593	83	15	models	model	NOUN
bracis-33593	83	16	approach	approach	NOUN
bracis-33593	83	17	,	,	PUNCT
bracis-33593	83	18	a	a	DET
bracis-33593	83	19	standard	standard	ADJ
bracis-33593	83	20	single	single	ADJ
bracis-33593	83	21	-	-	PUNCT
bracis-33593	83	22	model	model	NOUN
bracis-33593	83	23	approach	approach	NOUN
bracis-33593	83	24	,	,	PUNCT
bracis-33593	83	25	or	or	CCONJ
bracis-33593	83	26	even	even	ADV
bracis-33593	83	27	with	with	ADP
bracis-33593	83	28	real	real	ADJ
bracis-33593	83	29	-	-	PUNCT
bracis-33593	83	30	world	world	NOUN
bracis-33593	83	31	models	model	NOUN
bracis-33593	83	32	.	.	PUNCT
bracis-33593	84	1	by	by	ADP
bracis-33593	84	2	integrating	integrate	VERB
bracis-33593	84	3	these	these	DET
bracis-33593	84	4	components	component	NOUN
bracis-33593	84	5	,	,	PUNCT
bracis-33593	84	6	we	we	PRON
bracis-33593	84	7	create	create	VERB
bracis-33593	84	8	a	a	DET
bracis-33593	84	9	comprehensive	comprehensive	ADJ
bracis-33593	84	10	framework	framework	NOUN
bracis-33593	84	11	for	for	ADP
bracis-33593	84	12	modeling	model	VERB
bracis-33593	84	13	segmented	segment	VERB
bracis-33593	84	14	dynamics	dynamic	NOUN
bracis-33593	84	15	,	,	PUNCT
bracis-33593	84	16	learning	learn	VERB
bracis-33593	84	17	multiple	multiple	ADJ
bracis-33593	84	18	models	model	NOUN
bracis-33593	84	19	,	,	PUNCT
bracis-33593	84	20	and	and	CCONJ
bracis-33593	84	21	conducting	conduct	VERB
bracis-33593	84	22	performance	performance	NOUN
bracis-33593	84	23	evaluations	evaluation	NOUN
bracis-33593	84	24	across	across	ADP
bracis-33593	84	25	different	different	ADJ
bracis-33593	84	26	approaches	approach	NOUN
bracis-33593	84	27	.	.	PUNCT
bracis-33593	85	1	this	this	DET
bracis-33593	85	2	approach	approach	NOUN
bracis-33593	85	3	enables	enable	VERB
bracis-33593	85	4	us	we	PRON
bracis-33593	85	5	to	to	PART
bracis-33593	85	6	explore	explore	VERB
bracis-33593	85	7	the	the	DET
bracis-33593	85	8	effectiveness	effectiveness	NOUN
bracis-33593	85	9	of	of	ADP
bracis-33593	85	10	our	our	PRON
bracis-33593	85	11	proposed	propose	VERB
bracis-33593	85	12	method	method	NOUN
bracis-33593	85	13	in	in	ADP
bracis-33593	85	14	improving	improve	VERB
bracis-33593	85	15	sample	sample	NOUN
bracis-33593	85	16	efficiency	efficiency	NOUN
bracis-33593	85	17	and	and	CCONJ
bracis-33593	85	18	generalization	generalization	NOUN
bracis-33593	85	19	in	in	ADP
bracis-33593	85	20	reinforcement	reinforcement	NOUN
bracis-33593	85	21	learning	learning	NOUN
bracis-33593	85	22	tasks	task	NOUN
bracis-33593	85	23	.	.	PUNCT
bracis-33593	86	1	fig	fig	NOUN
bracis-33593	86	2	.	.	PUNCT
bracis-33593	87	1	1	1	X
bracis-33593	87	2	.	.	X
bracis-33593	87	3	modules	module	NOUN
bracis-33593	87	4	that	that	PRON
bracis-33593	87	5	compose	compose	VERB
bracis-33593	87	6	a	a	DET
bracis-33593	87	7	k	k	NOUN
bracis-33593	87	8	-	-	PUNCT
bracis-33593	87	9	model	model	NOUN
bracis-33593	87	10	’s	’s	PART
bracis-33593	87	11	agent	agent	NOUN
bracis-33593	87	12	that	that	PRON
bracis-33593	87	13	interacts	interact	VERB
bracis-33593	87	14	with	with	ADP
bracis-33593	87	15	the	the	DET
bracis-33593	87	16	environment	environment	NOUN
bracis-33593	87	17	full	full	ADJ
bracis-33593	87	18	size	size	NOUN
bracis-33593	87	19	image	image	NOUN
bracis-33593	87	20	figure	figure	NOUN
bracis-33593	87	21	 	 	SPACE
bracis-33593	87	22	1	1	NUM
bracis-33593	87	23	shows	show	VERB
bracis-33593	87	24	a	a	DET
bracis-33593	87	25	x	x	NOUN
bracis-33593	87	26	-	-	NOUN
bracis-33593	87	27	ray	ray	NOUN
bracis-33593	87	28	of	of	ADP
bracis-33593	87	29	the	the	DET
bracis-33593	87	30	proposed	propose	VERB
bracis-33593	87	31	framework	framework	NOUN
bracis-33593	87	32	that	that	PRON
bracis-33593	87	33	brings	bring	VERB
bracis-33593	87	34	k	k	PROPN
bracis-33593	87	35	parameters	parameter	NOUN
bracis-33593	87	36	that	that	PRON
bracis-33593	87	37	are	be	AUX
bracis-33593	87	38	updated	update	VERB
bracis-33593	87	39	with	with	ADP
bracis-33593	87	40	likelihood	likelihood	NOUN
bracis-33593	87	41	estimation	estimation	NOUN
bracis-33593	87	42	based	base	VERB
bracis-33593	87	43	on	on	ADP
bracis-33593	87	44	experienced	experienced	ADJ
bracis-33593	87	45	transitions	transition	NOUN
bracis-33593	87	46	.	.	PUNCT
bracis-33593	88	1	we	we	PRON
bracis-33593	88	2	can	can	AUX
bracis-33593	88	3	see	see	VERB
bracis-33593	88	4	above	above	ADP
bracis-33593	88	5	the	the	DET
bracis-33593	88	6	dashed	dash	VERB
bracis-33593	88	7	line	line	NOUN
bracis-33593	88	8	the	the	DET
bracis-33593	88	9	components	component	NOUN
bracis-33593	88	10	that	that	PRON
bracis-33593	88	11	are	be	AUX
bracis-33593	88	12	activated	activate	VERB
bracis-33593	88	13	every	every	DET
bracis-33593	88	14	time	time	NOUN
bracis-33593	88	15	step	step	NOUN
bracis-33593	88	16	,	,	PUNCT
bracis-33593	88	17	like	like	ADP
bracis-33593	88	18	the	the	DET
bracis-33593	88	19	memory	memory	NOUN
bracis-33593	88	20	that	that	SCONJ
bracis-33593	88	21	every	every	DET
bracis-33593	88	22	interaction	interaction	NOUN
bracis-33593	88	23	receives	receive	VERB
bracis-33593	88	24	new	new	ADJ
bracis-33593	88	25	data	datum	NOUN
bracis-33593	88	26	to	to	PART
bracis-33593	88	27	be	be	AUX
bracis-33593	88	28	stored	store	VERB
bracis-33593	88	29	and	and	CCONJ
bracis-33593	88	30	the	the	DET
bracis-33593	88	31	current	current	ADJ
bracis-33593	88	32	executing	executing	NOUN
bracis-33593	88	33	plan	plan	NOUN
bracis-33593	88	34	that	that	PRON
bracis-33593	88	35	is	be	AUX
bracis-33593	88	36	driving	drive	VERB
bracis-33593	88	37	the	the	DET
bracis-33593	88	38	decisions	decision	NOUN
bracis-33593	88	39	the	the	DET
bracis-33593	88	40	agent	agent	NOUN
bracis-33593	88	41	is	be	AUX
bracis-33593	88	42	taking	take	VERB
bracis-33593	88	43	every	every	DET
bracis-33593	88	44	step	step	NOUN
bracis-33593	88	45	.	.	PUNCT
bracis-33593	89	1	below	below	ADP
bracis-33593	89	2	the	the	DET
bracis-33593	89	3	dashed	dash	VERB
bracis-33593	89	4	line	line	NOUN
bracis-33593	89	5	are	be	AUX
bracis-33593	89	6	the	the	DET
bracis-33593	89	7	components	component	NOUN
bracis-33593	89	8	that	that	PRON
bracis-33593	89	9	are	be	AUX
bracis-33593	89	10	not	not	PART
bracis-33593	89	11	activated	activate	VERB
bracis-33593	89	12	every	every	DET
bracis-33593	89	13	time	time	NOUN
bracis-33593	89	14	step	step	NOUN
bracis-33593	89	15	like	like	ADP
bracis-33593	89	16	the	the	DET
bracis-33593	89	17	planner	planner	NOUN
bracis-33593	89	18	,	,	PUNCT
bracis-33593	89	19	that	that	PRON
bracis-33593	89	20	uses	use	VERB
bracis-33593	89	21	the	the	DET
bracis-33593	89	22	models	model	NOUN
bracis-33593	89	23	to	to	PART
bracis-33593	89	24	generate	generate	VERB
bracis-33593	89	25	a	a	DET
bracis-33593	89	26	new	new	ADJ
bracis-33593	89	27	plan	plan	NOUN
bracis-33593	89	28	,	,	PUNCT
bracis-33593	89	29	and	and	CCONJ
bracis-33593	89	30	the	the	DET
bracis-33593	89	31	estimator	estimator	NOUN
bracis-33593	89	32	that	that	PRON
bracis-33593	89	33	relies	rely	VERB
bracis-33593	89	34	on	on	ADP
bracis-33593	89	35	the	the	DET
bracis-33593	89	36	memory	memory	NOUN
bracis-33593	89	37	containing	contain	VERB
bracis-33593	89	38	new	new	ADJ
bracis-33593	89	39	observed	observe	VERB
bracis-33593	89	40	transitions	transition	NOUN
bracis-33593	89	41	together	together	ADV
bracis-33593	89	42	with	with	ADP
bracis-33593	89	43	past	past	ADJ
bracis-33593	89	44	experiences	experience	NOUN
bracis-33593	89	45	,	,	PUNCT
bracis-33593	89	46	to	to	PART
bracis-33593	89	47	update	update	VERB
bracis-33593	89	48	the	the	DET
bracis-33593	89	49	parameters	parameter	NOUN
bracis-33593	89	50	of	of	ADP
bracis-33593	89	51	the	the	DET
bracis-33593	89	52	models	model	NOUN
bracis-33593	89	53	.	.	PUNCT
bracis-33593	90	1	4.1	4.1	NUM
bracis-33593	90	2	likelihood	likelihood	NOUN
bracis-33593	90	3	estimator	estimator	NOUN
bracis-33593	90	4	drawing	drawing	NOUN
bracis-33593	90	5	inspiration	inspiration	NOUN
bracis-33593	90	6	from	from	ADP
bracis-33593	90	7	the	the	DET
bracis-33593	90	8	likelihood	likelihood	NOUN
bracis-33593	90	9	methodology	methodology	NOUN
bracis-33593	90	10	used	use	VERB
bracis-33593	90	11	to	to	PART
bracis-33593	90	12	find	find	VERB
bracis-33593	90	13	deep	deep	ADJ
bracis-33593	90	14	options	option	NOUN
bracis-33593	90	15	[	[	X
bracis-33593	90	16	4	4	NUM
bracis-33593	90	17	]	]	PUNCT
bracis-33593	90	18	,	,	PUNCT
bracis-33593	90	19	we	we	PRON
bracis-33593	90	20	devise	devise	VERB
bracis-33593	90	21	an	an	DET
bracis-33593	90	22	estimation	estimation	NOUN
bracis-33593	90	23	approach	approach	NOUN
bracis-33593	90	24	for	for	ADP
bracis-33593	90	25	all	all	DET
bracis-33593	90	26	parameters	parameter	NOUN
bracis-33593	90	27	\(\vartheta	\(\vartheta	ADJ
bracis-33593	90	28	\	\	PROPN
bracis-33593	90	29	)	)	PUNCT
bracis-33593	90	30	within	within	ADP
bracis-33593	90	31	our	our	PRON
bracis-33593	90	32	framework	framework	NOUN
bracis-33593	90	33	.	.	PUNCT
bracis-33593	91	1	for	for	ADP
bracis-33593	91	2	the	the	DET
bracis-33593	91	3	options	option	NOUN
bracis-33593	91	4	it	it	PRON
bracis-33593	91	5	was	be	AUX
bracis-33593	91	6	utilized	utilize	VERB
bracis-33593	91	7	the	the	DET
bracis-33593	91	8	likelihood	likelihood	NOUN
bracis-33593	91	9	in	in	ADP
bracis-33593	91	10	a	a	DET
bracis-33593	91	11	gradient	gradient	NOUN
bracis-33593	91	12	-	-	PUNCT
bracis-33593	91	13	based	base	VERB
bracis-33593	91	14	expectation	expectation	NOUN
bracis-33593	91	15	-	-	PUNCT
bracis-33593	91	16	maximization	maximization	NOUN
bracis-33593	91	17	algorithm	algorithm	NOUN
bracis-33593	91	18	to	to	PART
bracis-33593	91	19	identify	identify	VERB
bracis-33593	91	20	the	the	DET
bracis-33593	91	21	optimal	optimal	ADJ
bracis-33593	91	22	parameterized	parameterized	ADJ
bracis-33593	91	23	options	option	NOUN
bracis-33593	91	24	from	from	ADP
bracis-33593	91	25	observed	observed	ADJ
bracis-33593	91	26	trajectories	trajectory	NOUN
bracis-33593	91	27	.	.	PUNCT
bracis-33593	92	1	leveraging	leverage	VERB
bracis-33593	92	2	this	this	DET
bracis-33593	92	3	insight	insight	NOUN
bracis-33593	92	4	,	,	PUNCT
bracis-33593	92	5	we	we	PRON
bracis-33593	92	6	develop	develop	VERB
bracis-33593	92	7	a	a	DET
bracis-33593	92	8	likelihood	likelihood	NOUN
bracis-33593	92	9	function	function	NOUN
bracis-33593	92	10	tailored	tailor	VERB
bracis-33593	92	11	to	to	ADP
bracis-33593	92	12	our	our	PRON
bracis-33593	92	13	parameters	parameter	NOUN
bracis-33593	92	14	\(\vartheta	\(\vartheta	VERB
bracis-33593	92	15	\	\	PROPN
bracis-33593	92	16	)	)	PUNCT
bracis-33593	92	17	.	.	PUNCT
bracis-33593	93	1	our	our	PRON
bracis-33593	93	2	likelihood	likelihood	NOUN
bracis-33593	93	3	function	function	NOUN
bracis-33593	93	4	captures	capture	VERB
bracis-33593	93	5	the	the	DET
bracis-33593	93	6	probability	probability	NOUN
bracis-33593	93	7	of	of	ADP
bracis-33593	93	8	observing	observe	VERB
bracis-33593	93	9	the	the	DET
bracis-33593	93	10	transitions	transition	NOUN
bracis-33593	93	11	in	in	ADP
bracis-33593	93	12	the	the	DET
bracis-33593	93	13	environment	environment	NOUN
bracis-33593	93	14	given	give	VERB
bracis-33593	93	15	the	the	DET
bracis-33593	93	16	current	current	ADJ
bracis-33593	93	17	parameters	parameter	NOUN
bracis-33593	93	18	\(\vartheta	\(\vartheta	VERB
bracis-33593	93	19	\	\	PROPN
bracis-33593	93	20	)	)	PUNCT
bracis-33593	93	21	.	.	PUNCT
bracis-33593	94	1	by	by	ADP
bracis-33593	94	2	maximizing	maximize	VERB
bracis-33593	94	3	this	this	DET
bracis-33593	94	4	likelihood	likelihood	NOUN
bracis-33593	94	5	function	function	NOUN
bracis-33593	94	6	,	,	PUNCT
bracis-33593	94	7	we	we	PRON
bracis-33593	94	8	aim	aim	VERB
bracis-33593	94	9	to	to	PART
bracis-33593	94	10	iteratively	iteratively	ADV
bracis-33593	94	11	update	update	VERB
bracis-33593	94	12	and	and	CCONJ
bracis-33593	94	13	refine	refine	VERB
bracis-33593	94	14	our	our	PRON
bracis-33593	94	15	parameter	parameter	NOUN
bracis-33593	94	16	estimates	estimate	NOUN
bracis-33593	94	17	to	to	PART
bracis-33593	94	18	better	well	ADV
bracis-33593	94	19	align	align	VERB
bracis-33593	94	20	with	with	ADP
bracis-33593	94	21	observed	observed	ADJ
bracis-33593	94	22	data	datum	NOUN
bracis-33593	94	23	.	.	PUNCT
bracis-33593	95	1	to	to	PART
bracis-33593	95	2	achieve	achieve	VERB
bracis-33593	95	3	this	this	PRON
bracis-33593	95	4	,	,	PUNCT
bracis-33593	95	5	we	we	PRON
bracis-33593	95	6	employ	employ	VERB
bracis-33593	95	7	a	a	DET
bracis-33593	95	8	stochastic	stochastic	ADJ
bracis-33593	95	9	gradient	gradient	NOUN
bracis-33593	95	10	ascent	ascent	NOUN
bracis-33593	95	11	algorithm	algorithm	NOUN
bracis-33593	95	12	,	,	PUNCT
bracis-33593	95	13	which	which	PRON
bracis-33593	95	14	iteratively	iteratively	ADV
bracis-33593	95	15	adjusts	adjust	VERB
bracis-33593	95	16	the	the	DET
bracis-33593	95	17	parameters	parameter	NOUN
bracis-33593	95	18	\(\vartheta	\(\vartheta	VERB
bracis-33593	95	19	\	\	PROPN
bracis-33593	95	20	)	)	PUNCT
bracis-33593	95	21	based	base	VERB
bracis-33593	95	22	on	on	ADP
bracis-33593	95	23	online	online	ADJ
bracis-33593	95	24	experience	experience	NOUN
bracis-33593	95	25	.	.	PUNCT
bracis-33593	96	1	by	by	ADP
bracis-33593	96	2	maximizing	maximize	VERB
bracis-33593	96	3	the	the	DET
bracis-33593	96	4	log	log	NOUN
bracis-33593	96	5	likelihood	likelihood	NOUN
bracis-33593	96	6	of	of	ADP
bracis-33593	96	7	the	the	DET
bracis-33593	96	8	observed	observe	VERB
bracis-33593	96	9	data	datum	NOUN
bracis-33593	96	10	,	,	PUNCT
bracis-33593	96	11	we	we	PRON
bracis-33593	96	12	can	can	AUX
bracis-33593	96	13	effectively	effectively	ADV
bracis-33593	96	14	update	update	VERB
bracis-33593	96	15	our	our	PRON
bracis-33593	96	16	parameter	parameter	NOUN
bracis-33593	96	17	estimates	estimate	NOUN
bracis-33593	96	18	in	in	ADP
bracis-33593	96	19	a	a	DET
bracis-33593	96	20	principled	principled	ADJ
bracis-33593	96	21	manner	manner	NOUN
bracis-33593	96	22	,	,	PUNCT
bracis-33593	96	23	leveraging	leverage	VERB
bracis-33593	96	24	the	the	DET
bracis-33593	96	25	information	information	NOUN
bracis-33593	96	26	gleaned	glean	VERB
bracis-33593	96	27	from	from	ADP
bracis-33593	96	28	interactions	interaction	NOUN
bracis-33593	96	29	with	with	ADP
bracis-33593	96	30	the	the	DET
bracis-33593	96	31	environment	environment	NOUN
bracis-33593	96	32	.	.	PUNCT
bracis-33593	97	1	this	this	DET
bracis-33593	97	2	approach	approach	NOUN
bracis-33593	97	3	allows	allow	VERB
bracis-33593	97	4	us	we	PRON
bracis-33593	97	5	to	to	PART
bracis-33593	97	6	adaptively	adaptively	ADV
bracis-33593	97	7	learn	learn	VERB
bracis-33593	97	8	and	and	CCONJ
bracis-33593	97	9	refine	refine	VERB
bracis-33593	97	10	the	the	DET
bracis-33593	97	11	parameters	parameter	NOUN
bracis-33593	97	12	\(\vartheta	\(\vartheta	VERB
bracis-33593	97	13	\	\	PROPN
bracis-33593	97	14	)	)	PUNCT
bracis-33593	97	15	over	over	ADP
bracis-33593	97	16	time	time	NOUN
bracis-33593	97	17	,	,	PUNCT
bracis-33593	97	18	thereby	thereby	ADV
bracis-33593	97	19	improving	improve	VERB
bracis-33593	97	20	the	the	DET
bracis-33593	97	21	accuracy	accuracy	NOUN
bracis-33593	97	22	and	and	CCONJ
bracis-33593	97	23	effectiveness	effectiveness	NOUN
bracis-33593	97	24	of	of	ADP
bracis-33593	97	25	our	our	PRON
bracis-33593	97	26	multi	multi	ADJ
bracis-33593	97	27	model	model	NOUN
bracis-33593	97	28	-	-	PUNCT
bracis-33593	97	29	based	base	VERB
bracis-33593	97	30	reinforcement	reinforcement	NOUN
bracis-33593	97	31	learning	learning	NOUN
bracis-33593	97	32	system	system	NOUN
bracis-33593	97	33	.	.	PUNCT
bracis-33593	98	1	by	by	ADP
bracis-33593	98	2	continuously	continuously	ADV
bracis-33593	98	3	updating	update	VERB
bracis-33593	98	4	our	our	PRON
bracis-33593	98	5	parameter	parameter	NOUN
bracis-33593	98	6	estimates	estimate	NOUN
bracis-33593	98	7	based	base	VERB
bracis-33593	98	8	on	on	ADP
bracis-33593	98	9	observed	observed	ADJ
bracis-33593	98	10	data	datum	NOUN
bracis-33593	98	11	,	,	PUNCT
bracis-33593	98	12	we	we	PRON
bracis-33593	98	13	can	can	AUX
bracis-33593	98	14	enhance	enhance	VERB
bracis-33593	98	15	the	the	DET
bracis-33593	98	16	performance	performance	NOUN
bracis-33593	98	17	and	and	CCONJ
bracis-33593	98	18	efficiency	efficiency	NOUN
bracis-33593	98	19	of	of	ADP
bracis-33593	98	20	our	our	PRON
bracis-33593	98	21	algorithm	algorithm	NOUN
bracis-33593	98	22	in	in	ADP
bracis-33593	98	23	learning	learn	VERB
bracis-33593	98	24	the	the	DET
bracis-33593	98	25	dynamics	dynamic	NOUN
bracis-33593	98	26	of	of	ADP
bracis-33593	98	27	segmented	segment	VERB
bracis-33593	98	28	environments	environment	NOUN
bracis-33593	98	29	and	and	CCONJ
bracis-33593	98	30	generating	generate	VERB
bracis-33593	98	31	optimal	optimal	ADJ
bracis-33593	98	32	plans	plan	NOUN
bracis-33593	98	33	for	for	ADP
bracis-33593	98	34	achieving	achieve	VERB
bracis-33593	98	35	desired	desire	VERB
bracis-33593	98	36	goals	goal	NOUN
bracis-33593	98	37	.	.	PUNCT
bracis-33593	99	1	the	the	DET
bracis-33593	99	2	likelihood	likelihood	NOUN
bracis-33593	99	3	function	function	NOUN
bracis-33593	99	4	for	for	ADP
bracis-33593	99	5	a	a	DET
bracis-33593	99	6	trajectory	trajectory	NOUN
bracis-33593	99	7	\(\xi	\(\xi	PUNCT
bracis-33593	99	8	=	=	PUNCT
bracis-33593	99	9	{	{	PUNCT
bracis-33593	99	10	s_1	s_1	NOUN
bracis-33593	99	11	,	,	PUNCT
bracis-33593	99	12	a_1	a_1	NOUN
bracis-33593	99	13	,	,	PUNCT
bracis-33593	99	14	s_2	s_2	PROPN
bracis-33593	99	15	,	,	PUNCT
bracis-33593	99	16	a_2	a_2	PROPN
bracis-33593	99	17	,	,	PUNCT
bracis-33593	99	18	...	...	PUNCT
bracis-33593	99	19	,	,	PUNCT
bracis-33593	99	20	s_t}\	s_t}\	NOUN
bracis-33593	99	21	)	)	PUNCT
bracis-33593	99	22	in	in	ADP
bracis-33593	99	23	a	a	DET
bracis-33593	99	24	standard	standard	ADJ
bracis-33593	99	25	mdp	mdp	NOUN
bracis-33593	99	26	with	with	ADP
bracis-33593	99	27	a	a	DET
bracis-33593	99	28	model	model	NOUN
bracis-33593	99	29	parameterized	parameterize	VERB
bracis-33593	99	30	by	by	ADP
bracis-33593	99	31	\(\vartheta	\(\vartheta	ADJ
bracis-33593	99	32	\	\	NOUN
bracis-33593	99	33	)	)	PUNCT
bracis-33593	99	34	is	be	AUX
bracis-33593	99	35	calculated	calculate	VERB
bracis-33593	99	36	using	use	VERB
bracis-33593	99	37	the	the	DET
bracis-33593	99	38	transition	transition	NOUN
bracis-33593	99	39	probabilities	probability	NOUN
bracis-33593	99	40	as	as	SCONJ
bracis-33593	99	41	follows	follow	VERB
bracis-33593	99	42	:	:	PUNCT
bracis-33593	99	43	$	$	SYM
bracis-33593	99	44	$	$	SYM
bracis-33593	99	45	\begin{aligned	\begin{aligne	VERB
bracis-33593	99	46	}	}	PUNCT
bracis-33593	99	47	\mathcal{{l}}[\xi	\mathcal{{l}}[\xi	NOUN
bracis-33593	99	48	|	|	ADV
bracis-33593	99	49	\vartheta	\vartheta	INTJ
bracis-33593	99	50	]	]	X
bracis-33593	99	51	=	=	PUNCT
bracis-33593	99	52	\sum	\sum	NOUN
bracis-33593	99	53	_	_	PUNCT
bracis-33593	99	54	{	{	PUNCT
bracis-33593	99	55	t=1}^{t-1	t=1}^{t-1	PROPN
bracis-33593	99	56	}	}	PUNCT
bracis-33593	99	57	{	{	PUNCT
bracis-33593	99	58	\log	\log	PROPN
bracis-33593	99	59	\	\	PROPN
bracis-33593	99	60	;	;	PUNCT
bracis-33593	99	61	\mathcal	\mathcal	X
bracis-33593	99	62	{	{	PUNCT
bracis-33593	99	63	p}_{\vartheta	p}_{\vartheta	PROPN
bracis-33593	99	64	}	}	PUNCT
bracis-33593	99	65	(	(	PUNCT
bracis-33593	99	66	s_{t+1}|s_t	s_{t+1}|s_t	ADJ
bracis-33593	99	67	,	,	PUNCT
bracis-33593	99	68	a_t	a_t	NUM
bracis-33593	99	69	)	)	PUNCT
bracis-33593	99	70	}	}	PUNCT
bracis-33593	99	71	\end{aligned}$$	\end{aligned}$$	X
bracis-33593	99	72	(	(	PUNCT
bracis-33593	99	73	1	1	NUM
bracis-33593	99	74	)	)	PUNCT
bracis-33593	99	75	to	to	PART
bracis-33593	99	76	adapt	adapt	VERB
bracis-33593	99	77	this	this	DET
bracis-33593	99	78	likelihood	likelihood	NOUN
bracis-33593	99	79	to	to	ADP
bracis-33593	99	80	the	the	DET
bracis-33593	99	81	context	context	NOUN
bracis-33593	99	82	of	of	ADP
bracis-33593	99	83	k	k	PROPN
bracis-33593	99	84	models	model	NOUN
bracis-33593	99	85	,	,	PUNCT
bracis-33593	99	86	we	we	PRON
bracis-33593	99	87	incorporate	incorporate	VERB
bracis-33593	99	88	the	the	DET
bracis-33593	99	89	\(\beta	\(\beta	PROPN
bracis-33593	99	90	\	\	NOUN
bracis-33593	99	91	)	)	PUNCT
bracis-33593	99	92	function	function	NOUN
bracis-33593	99	93	and	and	CCONJ
bracis-33593	99	94	account	account	VERB
bracis-33593	99	95	for	for	ADP
bracis-33593	99	96	all	all	DET
bracis-33593	99	97	\(\mathcal	\(\mathcal	ADJ
bracis-33593	99	98	{	{	PUNCT
bracis-33593	99	99	p}_{\theta	p}_{\theta	NOUN
bracis-33593	99	100	}	}	PUNCT
bracis-33593	99	101	\	\	NOUN
bracis-33593	99	102	)	)	PUNCT
bracis-33593	99	103	as	as	SCONJ
bracis-33593	99	104	follows	follow	VERB
bracis-33593	99	105	:	:	PUNCT
bracis-33593	99	106	$	$	SYM
bracis-33593	99	107	$	$	SYM
bracis-33593	99	108	\begin{aligned	\begin{aligne	VERB
bracis-33593	99	109	}	}	PUNCT
bracis-33593	99	110	\mathcal	\mathcal	ADJ
bracis-33593	99	111	{	{	PUNCT
bracis-33593	99	112	p}_{\vartheta	p}_{\vartheta	NOUN
bracis-33593	99	113	}	}	PUNCT
bracis-33593	99	114	(	(	PUNCT
bracis-33593	99	115	s_{t+1}|s_t	s_{t+1}|s_t	ADJ
bracis-33593	99	116	,	,	PUNCT
bracis-33593	99	117	a_t	a_t	PUNCT
bracis-33593	99	118	)	)	PUNCT
bracis-33593	99	119	=	=	SYM
bracis-33593	99	120	\sum	\sum	NOUN
bracis-33593	99	121	_	_	PUNCT
bracis-33593	99	122	{	{	PUNCT
bracis-33593	99	123	\theta	\theta	PROPN
bracis-33593	99	124	\in	\in	PROPN
bracis-33593	99	125	\vartheta	\vartheta	PROPN
bracis-33593	99	126	}	}	PUNCT
bracis-33593	99	127	\mathcal	\mathcal	ADJ
bracis-33593	99	128	{	{	PUNCT
bracis-33593	99	129	p}_{\theta	p}_{\theta	NOUN
bracis-33593	99	130	}	}	PUNCT
bracis-33593	99	131	(	(	PUNCT
bracis-33593	99	132	s_{t+1}|	s_{t+1}|	PROPN
bracis-33593	99	133	s_t	s_t	PROPN
bracis-33593	99	134	,	,	PUNCT
bracis-33593	99	135	a_t	a_t	NUM
bracis-33593	99	136	)	)	PUNCT
bracis-33593	99	137	\beta	\beta	PROPN
bracis-33593	99	138	_	_	PUNCT
bracis-33593	99	139	{	{	PUNCT
bracis-33593	99	140	\theta	\theta	PROPN
bracis-33593	99	141	}	}	PUNCT
bracis-33593	99	142	(	(	PUNCT
bracis-33593	99	143	s_t	s_t	NOUN
bracis-33593	99	144	)	)	PUNCT
bracis-33593	99	145	\end{aligned}$$	\end{aligned}$$	X
bracis-33593	99	146	(	(	PUNCT
bracis-33593	99	147	2	2	X
bracis-33593	99	148	)	)	PUNCT
bracis-33593	99	149	we	we	PRON
bracis-33593	99	150	can	can	AUX
bracis-33593	99	151	use	use	VERB
bracis-33593	99	152	this	this	DET
bracis-33593	99	153	equation	equation	NOUN
bracis-33593	99	154	in	in	ADP
bracis-33593	99	155	the	the	DET
bracis-33593	99	156	traditional	traditional	ADJ
bracis-33593	99	157	likelihood	likelihood	NOUN
bracis-33593	99	158	equation	equation	NOUN
bracis-33593	99	159	.	.	PUNCT
bracis-33593	100	1	by	by	ADP
bracis-33593	100	2	iteratively	iteratively	ADV
bracis-33593	100	3	applying	apply	VERB
bracis-33593	100	4	stochastic	stochastic	ADJ
bracis-33593	100	5	gradient	gradient	ADJ
bracis-33593	100	6	descent	descent	NOUN
bracis-33593	100	7	on	on	ADP
bracis-33593	100	8	the	the	DET
bracis-33593	100	9	negative	negative	ADJ
bracis-33593	100	10	log	log	NOUN
bracis-33593	100	11	likelihood	likelihood	NOUN
bracis-33593	100	12	,	,	PUNCT
bracis-33593	100	13	we	we	PRON
bracis-33593	100	14	aim	aim	VERB
bracis-33593	100	15	to	to	PART
bracis-33593	100	16	optimize	optimize	VERB
bracis-33593	100	17	the	the	DET
bracis-33593	100	18	parameters	parameter	NOUN
bracis-33593	100	19	\(\theta	\(\theta	NOUN
bracis-33593	100	20	\	\	PROPN
bracis-33593	100	21	):	):	PUNCT
bracis-33593	100	22	$	$	SYM
bracis-33593	100	23	$	$	SYM
bracis-33593	100	24	\begin{aligned	\begin{aligne	VERB
bracis-33593	100	25	}	}	PUNCT
bracis-33593	100	26	\vartheta	\vartheta	PROPN
bracis-33593	100	27	=	=	PUNCT
bracis-33593	101	1	\vartheta	\vartheta	PROPN
bracis-33593	101	2	+	+	NUM
bracis-33593	101	3	\eta	\eta	PROPN
bracis-33593	101	4	\nabla	\nabla	PROPN
bracis-33593	101	5	\mathcal{{l}}[\xi	\mathcal{{l}}[\xi	PROPN
bracis-33593	101	6	|\vartheta	|\vartheta	PROPN
bracis-33593	101	7	]	]	PUNCT
bracis-33593	101	8	\alpha	\alpha	ADP
bracis-33593	101	9	\vardelta	\vardelta	NOUN
bracis-33593	101	10	\vartheta	\vartheta	PROPN
bracis-33593	101	11	\end{aligned}$$	\end{aligned}$$	X
bracis-33593	101	12	(	(	PUNCT
bracis-33593	101	13	3	3	NUM
bracis-33593	101	14	)	)	PUNCT
bracis-33593	101	15	where	where	SCONJ
bracis-33593	101	16	\(\eta	\(\eta	NOUN
bracis-33593	101	17	\	\	NOUN
bracis-33593	101	18	)	)	PUNCT
bracis-33593	101	19	represents	represent	VERB
bracis-33593	101	20	the	the	DET
bracis-33593	101	21	learning	learning	NOUN
bracis-33593	101	22	rate	rate	NOUN
bracis-33593	101	23	and	and	CCONJ
bracis-33593	101	24	\(\alpha	\(\alpha	X
bracis-33593	101	25	\	\	NOUN
bracis-33593	101	26	)	)	PUNCT
bracis-33593	101	27	is	be	AUX
bracis-33593	101	28	the	the	DET
bracis-33593	101	29	decay	decay	NOUN
bracis-33593	101	30	factor	factor	NOUN
bracis-33593	101	31	for	for	ADP
bracis-33593	101	32	momentum	momentum	NOUN
bracis-33593	101	33	.	.	PUNCT
bracis-33593	102	1	this	this	DET
bracis-33593	102	2	iterative	iterative	NOUN
bracis-33593	102	3	process	process	NOUN
bracis-33593	102	4	allows	allow	VERB
bracis-33593	102	5	us	we	PRON
bracis-33593	102	6	to	to	PART
bracis-33593	102	7	update	update	VERB
bracis-33593	102	8	the	the	DET
bracis-33593	102	9	parameters	parameter	NOUN
bracis-33593	102	10	\(\vartheta	\(\vartheta	VERB
bracis-33593	102	11	\	\	PROPN
bracis-33593	102	12	)	)	PUNCT
bracis-33593	102	13	in	in	ADP
bracis-33593	102	14	a	a	DET
bracis-33593	102	15	principled	principled	ADJ
bracis-33593	102	16	manner	manner	NOUN
bracis-33593	102	17	,	,	PUNCT
bracis-33593	102	18	leveraging	leverage	VERB
bracis-33593	102	19	the	the	DET
bracis-33593	102	20	observed	observed	ADJ
bracis-33593	102	21	trajectory	trajectory	NOUN
bracis-33593	102	22	\(\xi	\(\xi	PUNCT
bracis-33593	102	23	\	\	PROPN
bracis-33593	102	24	)	)	PUNCT
bracis-33593	102	25	to	to	PART
bracis-33593	102	26	refine	refine	VERB
bracis-33593	102	27	our	our	PRON
bracis-33593	102	28	model	model	NOUN
bracis-33593	102	29	and	and	CCONJ
bracis-33593	102	30	improve	improve	VERB
bracis-33593	102	31	its	its	PRON
bracis-33593	102	32	accuracy	accuracy	NOUN
bracis-33593	102	33	in	in	ADP
bracis-33593	102	34	capturing	capture	VERB
bracis-33593	102	35	the	the	DET
bracis-33593	102	36	dynamics	dynamic	NOUN
bracis-33593	102	37	of	of	ADP
bracis-33593	102	38	the	the	DET
bracis-33593	102	39	environment	environment	NOUN
bracis-33593	102	40	.	.	PUNCT
bracis-33593	103	1	4.2	4.2	NUM
bracis-33593	103	2	planner	planner	NOUN
bracis-33593	103	3	with	with	ADP
bracis-33593	103	4	the	the	DET
bracis-33593	103	5	established	establish	VERB
bracis-33593	103	6	models	model	NOUN
bracis-33593	103	7	we	we	PRON
bracis-33593	103	8	have	have	AUX
bracis-33593	103	9	chosen	choose	VERB
bracis-33593	103	10	a	a	DET
bracis-33593	103	11	planning	planning	NOUN
bracis-33593	103	12	approach	approach	NOUN
bracis-33593	103	13	for	for	ADP
bracis-33593	103	14	obtaining	obtain	VERB
bracis-33593	103	15	value	value	NOUN
bracis-33593	103	16	of	of	ADP
bracis-33593	103	17	the	the	DET
bracis-33593	103	18	representation	representation	NOUN
bracis-33593	103	19	of	of	ADP
bracis-33593	103	20	the	the	DET
bracis-33593	103	21	world	world	NOUN
bracis-33593	103	22	,	,	PUNCT
bracis-33593	103	23	inspired	inspire	VERB
bracis-33593	103	24	by	by	ADP
bracis-33593	103	25	the	the	DET
bracis-33593	103	26	implementation	implementation	NOUN
bracis-33593	103	27	rrt	rrt	NOUN
bracis-33593	103	28	based	base	VERB
bracis-33593	103	29	algorithms	algorithm	NOUN
bracis-33593	103	30	with	with	ADP
bracis-33593	103	31	rl	rl	PROPN
bracis-33593	103	32	[	[	X
bracis-33593	103	33	9	9	NUM
bracis-33593	103	34	]	]	PUNCT
bracis-33593	103	35	for	for	ADP
bracis-33593	103	36	optimizing	optimize	VERB
bracis-33593	103	37	policies	policy	NOUN
bracis-33593	103	38	to	to	PART
bracis-33593	103	39	find	find	VERB
bracis-33593	103	40	a	a	DET
bracis-33593	103	41	path	path	NOUN
bracis-33593	103	42	to	to	ADP
bracis-33593	103	43	the	the	DET
bracis-33593	103	44	goal	goal	NOUN
bracis-33593	103	45	state	state	NOUN
bracis-33593	103	46	,	,	PUNCT
bracis-33593	103	47	we	we	PRON
bracis-33593	103	48	decides	decide	VERB
bracis-33593	103	49	to	to	PART
bracis-33593	103	50	implement	implement	VERB
bracis-33593	103	51	our	our	PRON
bracis-33593	103	52	modified	modify	VERB
bracis-33593	103	53	version	version	NOUN
bracis-33593	103	54	of	of	ADP
bracis-33593	103	55	the	the	DET
bracis-33593	103	56	rrt	rrt	ADJ
bracis-33593	103	57	algorithm	algorithm	NOUN
bracis-33593	103	58	that	that	PRON
bracis-33593	103	59	makes	make	VERB
bracis-33593	103	60	use	use	NOUN
bracis-33593	103	61	of	of	ADP
bracis-33593	103	62	a	a	DET
bracis-33593	103	63	model	model	NOUN
bracis-33593	103	64	to	to	PART
bracis-33593	103	65	come	come	VERB
bracis-33593	103	66	with	with	ADP
bracis-33593	103	67	a	a	DET
bracis-33593	103	68	plan	plan	NOUN
bracis-33593	103	69	for	for	ADP
bracis-33593	103	70	reaching	reach	VERB
bracis-33593	103	71	the	the	DET
bracis-33593	103	72	goal	goal	NOUN
bracis-33593	103	73	.	.	PUNCT
bracis-33593	104	1	the	the	DET
bracis-33593	104	2	implementation	implementation	NOUN
bracis-33593	104	3	builds	build	VERB
bracis-33593	104	4	a	a	DET
bracis-33593	104	5	tree	tree	NOUN
bracis-33593	104	6	of	of	ADP
bracis-33593	104	7	state	state	NOUN
bracis-33593	104	8	transitions	transition	NOUN
bracis-33593	104	9	based	base	VERB
bracis-33593	104	10	on	on	ADP
bracis-33593	104	11	the	the	DET
bracis-33593	104	12	execution	execution	NOUN
bracis-33593	104	13	of	of	ADP
bracis-33593	104	14	actions	action	NOUN
bracis-33593	104	15	on	on	ADP
bracis-33593	104	16	a	a	DET
bracis-33593	104	17	model	model	NOUN
bracis-33593	104	18	passed	pass	VERB
bracis-33593	104	19	as	as	ADP
bracis-33593	104	20	input	input	NOUN
bracis-33593	104	21	,	,	PUNCT
bracis-33593	104	22	so	so	CCONJ
bracis-33593	104	23	the	the	DET
bracis-33593	104	24	model	model	NOUN
bracis-33593	104	25	can	can	AUX
bracis-33593	104	26	generate	generate	VERB
bracis-33593	104	27	synthetic	synthetic	ADJ
bracis-33593	104	28	transitions	transition	NOUN
bracis-33593	104	29	that	that	PRON
bracis-33593	104	30	are	be	AUX
bracis-33593	104	31	incorporated	incorporate	VERB
bracis-33593	104	32	to	to	ADP
bracis-33593	104	33	the	the	DET
bracis-33593	104	34	tree	tree	NOUN
bracis-33593	104	35	plan	plan	NOUN
bracis-33593	104	36	,	,	PUNCT
bracis-33593	104	37	when	when	SCONJ
bracis-33593	104	38	completed	complete	VERB
bracis-33593	104	39	the	the	DET
bracis-33593	104	40	tree	tree	NOUN
bracis-33593	104	41	starts	start	VERB
bracis-33593	104	42	from	from	ADP
bracis-33593	104	43	a	a	DET
bracis-33593	104	44	start	start	NOUN
bracis-33593	104	45	state	state	NOUN
bracis-33593	104	46	,	,	PUNCT
bracis-33593	104	47	and	and	CCONJ
bracis-33593	104	48	have	have	VERB
bracis-33593	104	49	a	a	DET
bracis-33593	104	50	depth	depth	NOUN
bracis-33593	104	51	of	of	ADP
bracis-33593	104	52	a	a	DET
bracis-33593	104	53	certain	certain	ADJ
bracis-33593	104	54	max	max	NOUN
bracis-33593	104	55	number	number	NOUN
bracis-33593	104	56	of	of	ADP
bracis-33593	104	57	steps	step	NOUN
bracis-33593	104	58	that	that	PRON
bracis-33593	104	59	we	we	PRON
bracis-33593	104	60	wish	wish	VERB
bracis-33593	104	61	the	the	DET
bracis-33593	104	62	plan	plan	NOUN
bracis-33593	104	63	to	to	PART
bracis-33593	104	64	have	have	AUX
bracis-33593	104	65	.	.	PUNCT
bracis-33593	105	1	the	the	DET
bracis-33593	105	2	algorithm	algorithm	NOUN
bracis-33593	105	3	 	 	SPACE
bracis-33593	105	4	1	1	NUM
bracis-33593	105	5	is	be	AUX
bracis-33593	105	6	responsible	responsible	ADJ
bracis-33593	105	7	for	for	ADP
bracis-33593	105	8	building	build	VERB
bracis-33593	105	9	this	this	DET
bracis-33593	105	10	plan	plan	NOUN
bracis-33593	105	11	tree	tree	NOUN
bracis-33593	105	12	.	.	PUNCT
bracis-33593	106	1	algorithm	algorithm	NOUN
bracis-33593	106	2	1	1	NUM
bracis-33593	106	3	.	.	PUNCT
bracis-33593	107	1	rrt	rrt	NOUN
bracis-33593	107	2	based	base	VERB
bracis-33593	107	3	planner	planner	NOUN
bracis-33593	107	4	full	full	ADJ
bracis-33593	107	5	size	size	NOUN
bracis-33593	107	6	image	image	NOUN
bracis-33593	107	7	once	once	SCONJ
bracis-33593	107	8	the	the	DET
bracis-33593	107	9	plan	plan	NOUN
bracis-33593	107	10	tree	tree	NOUN
bracis-33593	107	11	is	be	AUX
bracis-33593	107	12	constructed	construct	VERB
bracis-33593	107	13	,	,	PUNCT
bracis-33593	107	14	we	we	PRON
bracis-33593	107	15	can	can	AUX
bracis-33593	107	16	extract	extract	VERB
bracis-33593	107	17	a	a	DET
bracis-33593	107	18	path	path	NOUN
bracis-33593	107	19	to	to	ADP
bracis-33593	107	20	the	the	DET
bracis-33593	107	21	goal	goal	NOUN
bracis-33593	107	22	state	state	NOUN
bracis-33593	107	23	or	or	CCONJ
bracis-33593	107	24	to	to	ADP
bracis-33593	107	25	the	the	DET
bracis-33593	107	26	state	state	NOUN
bracis-33593	107	27	within	within	ADP
bracis-33593	107	28	the	the	DET
bracis-33593	107	29	tree	tree	NOUN
bracis-33593	107	30	that	that	PRON
bracis-33593	107	31	is	be	AUX
bracis-33593	107	32	closest	close	ADJ
bracis-33593	107	33	to	to	ADP
bracis-33593	107	34	the	the	DET
bracis-33593	107	35	goal	goal	NOUN
bracis-33593	107	36	.	.	PUNCT
bracis-33593	108	1	to	to	PART
bracis-33593	108	2	accomplish	accomplish	VERB
bracis-33593	108	3	this	this	PRON
bracis-33593	108	4	,	,	PUNCT
bracis-33593	108	5	we	we	PRON
bracis-33593	108	6	only	only	ADV
bracis-33593	108	7	require	require	VERB
bracis-33593	108	8	a	a	DET
bracis-33593	108	9	function	function	NOUN
bracis-33593	108	10	to	to	PART
bracis-33593	108	11	calculate	calculate	VERB
bracis-33593	108	12	the	the	DET
bracis-33593	108	13	relative	relative	ADJ
bracis-33593	108	14	distance	distance	NOUN
bracis-33593	108	15	between	between	ADP
bracis-33593	108	16	states	state	NOUN
bracis-33593	108	17	.	.	PUNCT
bracis-33593	109	1	algorithm	algorithm	NOUN
bracis-33593	109	2	 	 	SPACE
bracis-33593	109	3	2	2	NUM
bracis-33593	109	4	provides	provide	VERB
bracis-33593	109	5	a	a	DET
bracis-33593	109	6	systematic	systematic	ADJ
bracis-33593	109	7	method	method	NOUN
bracis-33593	109	8	for	for	ADP
bracis-33593	109	9	extracting	extract	VERB
bracis-33593	109	10	a	a	DET
bracis-33593	109	11	path	path	NOUN
bracis-33593	109	12	from	from	ADP
bracis-33593	109	13	the	the	DET
bracis-33593	109	14	input	input	NOUN
bracis-33593	109	15	tree	tree	NOUN
bracis-33593	109	16	to	to	ADP
bracis-33593	109	17	either	either	CCONJ
bracis-33593	109	18	the	the	DET
bracis-33593	109	19	goal	goal	NOUN
bracis-33593	109	20	state	state	NOUN
bracis-33593	109	21	or	or	CCONJ
bracis-33593	109	22	the	the	DET
bracis-33593	109	23	closest	close	ADJ
bracis-33593	109	24	state	state	NOUN
bracis-33593	109	25	found	find	VERB
bracis-33593	109	26	using	use	VERB
bracis-33593	109	27	the	the	DET
bracis-33593	109	28	tree.nearest	tree.near	ADJ
bracis-33593	109	29	function	function	NOUN
bracis-33593	109	30	.	.	PUNCT
bracis-33593	110	1	algorithm	algorithm	NOUN
bracis-33593	110	2	2	2	NUM
bracis-33593	110	3	.	.	PUNCT
bracis-33593	111	1	get	get	VERB
bracis-33593	111	2	path	path	NOUN
bracis-33593	111	3	full	full	ADJ
bracis-33593	111	4	size	size	NOUN
bracis-33593	111	5	image	image	NOUN
bracis-33593	111	6	the	the	DET
bracis-33593	111	7	tree.nearest	tree.near	ADJ
bracis-33593	111	8	function	function	NOUN
bracis-33593	111	9	plays	play	VERB
bracis-33593	111	10	a	a	DET
bracis-33593	111	11	crucial	crucial	ADJ
bracis-33593	111	12	role	role	NOUN
bracis-33593	111	13	in	in	ADP
bracis-33593	111	14	this	this	DET
bracis-33593	111	15	process	process	NOUN
bracis-33593	111	16	by	by	ADP
bracis-33593	111	17	identifying	identify	VERB
bracis-33593	111	18	the	the	DET
bracis-33593	111	19	closest	close	ADJ
bracis-33593	111	20	state	state	NOUN
bracis-33593	111	21	to	to	ADP
bracis-33593	111	22	the	the	DET
bracis-33593	111	23	goal	goal	NOUN
bracis-33593	111	24	within	within	ADP
bracis-33593	111	25	the	the	DET
bracis-33593	111	26	plan	plan	NOUN
bracis-33593	111	27	tree	tree	NOUN
bracis-33593	111	28	.	.	PUNCT
bracis-33593	112	1	in	in	ADP
bracis-33593	112	2	this	this	DET
bracis-33593	112	3	case	case	NOUN
bracis-33593	112	4	it	it	PRON
bracis-33593	112	5	is	be	AUX
bracis-33593	112	6	used	use	VERB
bracis-33593	112	7	the	the	DET
bracis-33593	112	8	euclidean	euclidean	ADJ
bracis-33593	112	9	distance	distance	NOUN
bracis-33593	112	10	to	to	PART
bracis-33593	112	11	compare	compare	VERB
bracis-33593	112	12	the	the	DET
bracis-33593	112	13	states	state	NOUN
bracis-33593	112	14	distances	distance	NOUN
bracis-33593	112	15	.	.	PUNCT
bracis-33593	113	1	once	once	ADV
bracis-33593	113	2	this	this	DET
bracis-33593	113	3	closest	close	ADJ
bracis-33593	113	4	state	state	NOUN
bracis-33593	113	5	is	be	AUX
bracis-33593	113	6	determined	determine	VERB
bracis-33593	113	7	,	,	PUNCT
bracis-33593	113	8	the	the	DET
bracis-33593	113	9	algorithm	algorithm	NOUN
bracis-33593	113	10	proceeds	proceed	VERB
bracis-33593	113	11	to	to	PART
bracis-33593	113	12	construct	construct	VERB
bracis-33593	113	13	a	a	DET
bracis-33593	113	14	path	path	NOUN
bracis-33593	113	15	from	from	ADP
bracis-33593	113	16	the	the	DET
bracis-33593	113	17	initial	initial	ADJ
bracis-33593	113	18	state	state	NOUN
bracis-33593	113	19	to	to	ADP
bracis-33593	113	20	either	either	CCONJ
bracis-33593	113	21	the	the	DET
bracis-33593	113	22	goal	goal	NOUN
bracis-33593	113	23	state	state	NOUN
bracis-33593	113	24	or	or	CCONJ
bracis-33593	113	25	the	the	DET
bracis-33593	113	26	identified	identify	VERB
bracis-33593	113	27	closest	close	ADJ
bracis-33593	113	28	state	state	NOUN
bracis-33593	113	29	.	.	PUNCT
bracis-33593	114	1	by	by	ADP
bracis-33593	114	2	leveraging	leverage	VERB
bracis-33593	114	3	algorithm	algorithm	NOUN
bracis-33593	114	4	 	 	SPACE
bracis-33593	114	5	2	2	NUM
bracis-33593	114	6	,	,	PUNCT
bracis-33593	114	7	the	the	DET
bracis-33593	114	8	agent	agent	NOUN
bracis-33593	114	9	can	can	AUX
bracis-33593	114	10	effectively	effectively	ADV
bracis-33593	114	11	obtain	obtain	VERB
bracis-33593	114	12	a	a	DET
bracis-33593	114	13	plan	plan	NOUN
bracis-33593	114	14	of	of	ADP
bracis-33593	114	15	actions	action	NOUN
bracis-33593	114	16	to	to	PART
bracis-33593	114	17	take	take	VERB
bracis-33593	114	18	.	.	PUNCT
bracis-33593	115	1	the	the	DET
bracis-33593	115	2	quality	quality	NOUN
bracis-33593	115	3	of	of	ADP
bracis-33593	115	4	plans	plan	NOUN
bracis-33593	115	5	generated	generate	VERB
bracis-33593	115	6	by	by	ADP
bracis-33593	115	7	the	the	DET
bracis-33593	115	8	algorithm	algorithm	NOUN
bracis-33593	115	9	is	be	AUX
bracis-33593	115	10	directly	directly	ADV
bracis-33593	115	11	influenced	influence	VERB
bracis-33593	115	12	by	by	ADP
bracis-33593	115	13	the	the	DET
bracis-33593	115	14	accuracy	accuracy	NOUN
bracis-33593	115	15	and	and	CCONJ
bracis-33593	115	16	fidelity	fidelity	NOUN
bracis-33593	115	17	of	of	ADP
bracis-33593	115	18	the	the	DET
bracis-33593	115	19	underlying	underlie	VERB
bracis-33593	115	20	models	model	NOUN
bracis-33593	115	21	used	use	VERB
bracis-33593	115	22	to	to	PART
bracis-33593	115	23	construct	construct	VERB
bracis-33593	115	24	them	they	PRON
bracis-33593	115	25	.	.	PUNCT
bracis-33593	116	1	initially	initially	ADV
bracis-33593	116	2	,	,	PUNCT
bracis-33593	116	3	when	when	SCONJ
bracis-33593	116	4	the	the	DET
bracis-33593	116	5	models	model	NOUN
bracis-33593	116	6	are	be	AUX
bracis-33593	116	7	far	far	ADV
bracis-33593	116	8	from	from	ADP
bracis-33593	116	9	reflecting	reflect	VERB
bracis-33593	116	10	the	the	DET
bracis-33593	116	11	real	real	ADJ
bracis-33593	116	12	transitions	transition	NOUN
bracis-33593	116	13	in	in	ADP
bracis-33593	116	14	the	the	DET
bracis-33593	116	15	environment	environment	NOUN
bracis-33593	116	16	,	,	PUNCT
bracis-33593	116	17	the	the	DET
bracis-33593	116	18	generated	generate	VERB
bracis-33593	116	19	plans	plan	NOUN
bracis-33593	116	20	may	may	AUX
bracis-33593	116	21	be	be	AUX
bracis-33593	116	22	misguided	misguided	ADJ
bracis-33593	116	23	or	or	CCONJ
bracis-33593	116	24	sub	sub	VERB
bracis-33593	116	25	optimal	optimal	ADJ
bracis-33593	116	26	.	.	PUNCT
bracis-33593	117	1	however	however	ADV
bracis-33593	117	2	,	,	PUNCT
bracis-33593	117	3	as	as	SCONJ
bracis-33593	117	4	the	the	DET
bracis-33593	117	5	agent	agent	NOUN
bracis-33593	117	6	gathers	gather	VERB
bracis-33593	117	7	more	more	ADJ
bracis-33593	117	8	experience	experience	NOUN
bracis-33593	117	9	and	and	CCONJ
bracis-33593	117	10	refines	refine	VERB
bracis-33593	117	11	its	its	PRON
bracis-33593	117	12	models	model	NOUN
bracis-33593	117	13	to	to	PART
bracis-33593	117	14	better	well	ADV
bracis-33593	117	15	reflect	reflect	VERB
bracis-33593	117	16	the	the	DET
bracis-33593	117	17	true	true	ADJ
bracis-33593	117	18	dynamics	dynamic	NOUN
bracis-33593	117	19	of	of	ADP
bracis-33593	117	20	the	the	DET
bracis-33593	117	21	environment	environment	NOUN
bracis-33593	117	22	,	,	PUNCT
bracis-33593	117	23	the	the	DET
bracis-33593	117	24	quality	quality	NOUN
bracis-33593	117	25	of	of	ADP
bracis-33593	117	26	the	the	DET
bracis-33593	117	27	generated	generate	VERB
bracis-33593	117	28	plans	plan	NOUN
bracis-33593	117	29	improves	improve	VERB
bracis-33593	117	30	significantly	significantly	ADV
bracis-33593	117	31	.	.	PUNCT
bracis-33593	118	1	this	this	DET
bracis-33593	118	2	highlights	highlight	VERB
bracis-33593	118	3	the	the	DET
bracis-33593	118	4	iterative	iterative	ADJ
bracis-33593	118	5	nature	nature	NOUN
bracis-33593	118	6	of	of	ADP
bracis-33593	118	7	model	model	NOUN
bracis-33593	118	8	-	-	PUNCT
bracis-33593	118	9	based	base	VERB
bracis-33593	118	10	reinforcement	reinforcement	NOUN
bracis-33593	118	11	learning	learning	NOUN
bracis-33593	118	12	,	,	PUNCT
bracis-33593	118	13	and	and	CCONJ
bracis-33593	118	14	works	work	VERB
bracis-33593	118	15	[	[	X
bracis-33593	118	16	14	14	NUM
bracis-33593	118	17	]	]	PUNCT
bracis-33593	118	18	have	have	AUX
bracis-33593	118	19	even	even	ADV
bracis-33593	118	20	shown	show	VERB
bracis-33593	118	21	that	that	SCONJ
bracis-33593	118	22	this	this	DET
bracis-33593	118	23	initial	initial	ADJ
bracis-33593	118	24	misguidance	misguidance	NOUN
bracis-33593	118	25	of	of	ADP
bracis-33593	118	26	poor	poor	ADJ
bracis-33593	118	27	models	model	NOUN
bracis-33593	118	28	could	could	AUX
bracis-33593	118	29	bring	bring	VERB
bracis-33593	118	30	valuable	valuable	ADJ
bracis-33593	118	31	exploration	exploration	NOUN
bracis-33593	118	32	experiences	experience	NOUN
bracis-33593	118	33	in	in	ADP
bracis-33593	118	34	some	some	DET
bracis-33593	118	35	specific	specific	ADJ
bracis-33593	118	36	problems	problem	NOUN
bracis-33593	118	37	.	.	PUNCT
bracis-33593	119	1	5	5	NUM
bracis-33593	119	2	applied	apply	VERB
bracis-33593	119	3	environment	environment	NOUN
bracis-33593	119	4	our	our	PRON
bracis-33593	119	5	challenge	challenge	NOUN
bracis-33593	119	6	lies	lie	VERB
bracis-33593	119	7	in	in	ADP
bracis-33593	119	8	simultaneously	simultaneously	ADV
bracis-33593	119	9	learning	learn	VERB
bracis-33593	119	10	the	the	DET
bracis-33593	119	11	k	k	PROPN
bracis-33593	119	12	parameters	parameters	PROPN
bracis-33593	119	13	\(\theta	\(\theta	NOUN
bracis-33593	119	14	\	\	PROPN
bracis-33593	119	15	)	)	PUNCT
bracis-33593	119	16	and	and	CCONJ
bracis-33593	119	17	the	the	DET
bracis-33593	119	18	probability	probability	NOUN
bracis-33593	119	19	distribution	distribution	NOUN
bracis-33593	119	20	given	give	VERB
bracis-33593	119	21	by	by	ADP
bracis-33593	119	22	\(\beta	\(\beta	X
bracis-33593	119	23	\	\	NOUN
bracis-33593	119	24	)	)	PUNCT
bracis-33593	119	25	across	across	ADP
bracis-33593	119	26	states	state	NOUN
bracis-33593	119	27	.	.	PUNCT
bracis-33593	120	1	to	to	PART
bracis-33593	120	2	address	address	VERB
bracis-33593	120	3	this	this	DET
bracis-33593	120	4	challenge	challenge	NOUN
bracis-33593	120	5	,	,	PUNCT
bracis-33593	120	6	we	we	PRON
bracis-33593	120	7	assume	assume	VERB
bracis-33593	120	8	that	that	SCONJ
bracis-33593	120	9	the	the	DET
bracis-33593	120	10	function	function	NOUN
bracis-33593	120	11	\(\mathcal	\(\mathcal	ADJ
bracis-33593	120	12	{	{	PUNCT
bracis-33593	120	13	p}\	p}\	NOUN
bracis-33593	120	14	)	)	PUNCT
bracis-33593	120	15	is	be	AUX
bracis-33593	120	16	based	base	VERB
bracis-33593	120	17	in	in	ADP
bracis-33593	120	18	a	a	DET
bracis-33593	120	19	gaussian	gaussian	NOUN
bracis-33593	120	20	parameterized	parameterize	VERB
bracis-33593	120	21	,	,	PUNCT
bracis-33593	120	22	with	with	ADP
bracis-33593	120	23	\(\theta	\(\theta	NOUN
bracis-33593	120	24	_	_	NOUN
bracis-33593	121	1	i	i	PRON
bracis-33593	121	2	=	=	PUNCT
bracis-33593	121	3	\{\tau	\{\tau	VERB
bracis-33593	121	4	_	_	PUNCT
bracis-33593	122	1	i	i	PRON
bracis-33593	122	2	,	,	PUNCT
bracis-33593	122	3	\sigma	\sigma	PROPN
bracis-33593	122	4	_	_	PUNCT
bracis-33593	122	5	i\}\	i\}\	ADV
bracis-33593	122	6	)	)	PUNCT
bracis-33593	122	7	representing	represent	VERB
bracis-33593	122	8	the	the	DET
bracis-33593	122	9	parameters	parameter	NOUN
bracis-33593	122	10	that	that	PRON
bracis-33593	122	11	will	will	AUX
bracis-33593	122	12	be	be	AUX
bracis-33593	122	13	used	use	VERB
bracis-33593	122	14	in	in	ADP
bracis-33593	122	15	the	the	DET
bracis-33593	122	16	transitions	transition	NOUN
bracis-33593	122	17	.	.	PUNCT
bracis-33593	123	1	to	to	PART
bracis-33593	123	2	create	create	VERB
bracis-33593	123	3	a	a	DET
bracis-33593	123	4	simulation	simulation	NOUN
bracis-33593	123	5	environment	environment	NOUN
bracis-33593	123	6	reflecting	reflect	VERB
bracis-33593	123	7	practical	practical	ADJ
bracis-33593	123	8	problems	problem	NOUN
bracis-33593	123	9	characterized	characterize	VERB
bracis-33593	123	10	by	by	ADP
bracis-33593	123	11	segmented	segment	VERB
bracis-33593	123	12	dynamics	dynamic	NOUN
bracis-33593	123	13	,	,	PUNCT
bracis-33593	123	14	we	we	PRON
bracis-33593	123	15	developed	develop	VERB
bracis-33593	123	16	a	a	DET
bracis-33593	123	17	two	two	NUM
bracis-33593	123	18	-	-	PUNCT
bracis-33593	123	19	dimensional	dimensional	ADJ
bracis-33593	123	20	space	space	NOUN
bracis-33593	123	21	where	where	SCONJ
bracis-33593	123	22	states	state	NOUN
bracis-33593	123	23	are	be	AUX
bracis-33593	123	24	represented	represent	VERB
bracis-33593	123	25	by	by	ADP
bracis-33593	123	26	x	x	SYM
bracis-33593	123	27	and	and	CCONJ
bracis-33593	123	28	y	y	PROPN
bracis-33593	123	29	coordinates	coordinate	NOUN
bracis-33593	123	30	specifying	specify	VERB
bracis-33593	123	31	\(\mathcal	\(\mathcal	ADJ
bracis-33593	123	32	{	{	PUNCT
bracis-33593	123	33	s	s	NOUN
bracis-33593	123	34	}	}	PUNCT
bracis-33593	123	35	\in	\in	ADJ
bracis-33593	123	36	\mathbb	\mathbb	PROPN
bracis-33593	123	37	{	{	PUNCT
bracis-33593	123	38	r}^2\	r}^2\	NOUN
bracis-33593	123	39	)	)	PUNCT
bracis-33593	123	40	.	.	PUNCT
bracis-33593	124	1	in	in	ADP
bracis-33593	124	2	this	this	DET
bracis-33593	124	3	environment	environment	NOUN
bracis-33593	124	4	,	,	PUNCT
bracis-33593	124	5	agents	agent	NOUN
bracis-33593	124	6	have	have	VERB
bracis-33593	124	7	four	four	NUM
bracis-33593	124	8	possible	possible	ADJ
bracis-33593	124	9	actions	action	NOUN
bracis-33593	124	10	corresponding	correspond	VERB
bracis-33593	124	11	to	to	ADP
bracis-33593	124	12	movement	movement	NOUN
bracis-33593	124	13	in	in	ADP
bracis-33593	124	14	the	the	DET
bracis-33593	124	15	cardinal	cardinal	ADJ
bracis-33593	124	16	directions	direction	NOUN
bracis-33593	124	17	then	then	ADV
bracis-33593	124	18	giving	give	VERB
bracis-33593	124	19	us	we	PRON
bracis-33593	124	20	\(\mathcal	\(\mathcal	ADJ
bracis-33593	124	21	{	{	PUNCT
bracis-33593	124	22	a	a	DET
bracis-33593	124	23	}	}	PUNCT
bracis-33593	124	24	\in	\in	PROPN
bracis-33593	124	25	\{\uparrow	\{\uparrow	ADV
bracis-33593	124	26	,	,	PUNCT
bracis-33593	124	27	\leftarrow	\leftarrow	PROPN
bracis-33593	124	28	,	,	PUNCT
bracis-33593	124	29	\downarrow	\downarrow	PROPN
bracis-33593	124	30	,	,	PUNCT
bracis-33593	124	31	\rightarrow	\rightarrow	PROPN
bracis-33593	124	32	\	\	ADJ
bracis-33593	124	33	}	}	PUNCT
bracis-33593	124	34	=	=	SYM
bracis-33593	124	35	\{(0,-1	\{(0,-1	NOUN
bracis-33593	124	36	)	)	PUNCT
bracis-33593	124	37	,	,	PUNCT
bracis-33593	124	38	(	(	PUNCT
bracis-33593	124	39	-1	-1	INTJ
bracis-33593	124	40	,	,	PUNCT
bracis-33593	124	41	0	0	NUM
bracis-33593	124	42	)	)	PUNCT
bracis-33593	124	43	,	,	PUNCT
bracis-33593	124	44	(	(	PUNCT
bracis-33593	124	45	0	0	NUM
bracis-33593	124	46	,	,	PUNCT
bracis-33593	124	47	1	1	NUM
bracis-33593	124	48	)	)	PUNCT
bracis-33593	124	49	,	,	PUNCT
bracis-33593	124	50	(	(	PUNCT
bracis-33593	124	51	1,0)\}\	1,0)\}\	NUM
bracis-33593	124	52	)	)	PUNCT
bracis-33593	124	53	.	.	PUNCT
bracis-33593	125	1	the	the	DET
bracis-33593	125	2	uncertainty	uncertainty	NOUN
bracis-33593	125	3	of	of	ADP
bracis-33593	125	4	the	the	DET
bracis-33593	125	5	agents	agent	NOUN
bracis-33593	125	6	’	’	PART
bracis-33593	125	7	movements	movement	NOUN
bracis-33593	125	8	,	,	PUNCT
bracis-33593	125	9	are	be	AUX
bracis-33593	125	10	formulated	formulate	VERB
bracis-33593	125	11	in	in	ADP
bracis-33593	125	12	transitions	transition	NOUN
bracis-33593	125	13	as	as	ADP
bracis-33593	125	14	linear	linear	ADJ
bracis-33593	125	15	transformations	transformation	NOUN
bracis-33593	125	16	of	of	ADP
bracis-33593	125	17	the	the	DET
bracis-33593	125	18	state	state	NOUN
bracis-33593	125	19	with	with	ADP
bracis-33593	125	20	respect	respect	NOUN
bracis-33593	125	21	to	to	ADP
bracis-33593	125	22	the	the	DET
bracis-33593	125	23	chosen	choose	VERB
bracis-33593	125	24	action	action	NOUN
bracis-33593	125	25	with	with	ADP
bracis-33593	125	26	a	a	DET
bracis-33593	125	27	random	random	ADJ
bracis-33593	125	28	factor	factor	NOUN
bracis-33593	125	29	drawn	draw	VERB
bracis-33593	125	30	from	from	ADP
bracis-33593	125	31	a	a	DET
bracis-33593	125	32	gaussian	gaussian	ADJ
bracis-33593	125	33	distribution	distribution	NOUN
bracis-33593	125	34	.	.	PUNCT
bracis-33593	126	1	notably	notably	ADV
bracis-33593	126	2	,	,	PUNCT
bracis-33593	126	3	the	the	DET
bracis-33593	126	4	parameters	parameter	NOUN
bracis-33593	126	5	of	of	ADP
bracis-33593	126	6	this	this	DET
bracis-33593	126	7	distribution	distribution	NOUN
bracis-33593	126	8	vary	vary	VERB
bracis-33593	126	9	across	across	ADP
bracis-33593	126	10	distinct	distinct	ADJ
bracis-33593	126	11	regions	region	NOUN
bracis-33593	126	12	of	of	ADP
bracis-33593	126	13	the	the	DET
bracis-33593	126	14	state	state	NOUN
bracis-33593	126	15	space	space	NOUN
bracis-33593	126	16	.	.	PUNCT
bracis-33593	127	1	so	so	ADV
bracis-33593	127	2	if	if	SCONJ
bracis-33593	127	3	we	we	PRON
bracis-33593	127	4	 	 	SPACE
bracis-33593	127	5	have	have	AUX
bracis-33593	127	6	k	k	PROPN
bracis-33593	127	7	regions	region	NOUN
bracis-33593	127	8	resulting	result	VERB
bracis-33593	127	9	 	 	SPACE
bracis-33593	127	10	state	state	NOUN
bracis-33593	127	11	\(s'\	\(s'\	PROPN
bracis-33593	127	12	)	)	PUNCT
bracis-33593	127	13	after	after	ADP
bracis-33593	127	14	executing	execute	VERB
bracis-33593	127	15	action	action	NOUN
bracis-33593	127	16	a	a	DET
bracis-33593	127	17	in	in	ADP
bracis-33593	127	18	 	 	SPACE
bracis-33593	127	19	state	state	NOUN
bracis-33593	127	20	s	s	VERB
bracis-33593	127	21	would	would	AUX
bracis-33593	127	22	 	 	SPACE
bracis-33593	127	23	be	be	AUX
bracis-33593	127	24	\(s	\(	NOUN
bracis-33593	127	25	'	'	PART
bracis-33593	127	26	\sim	\sim	NOUN
bracis-33593	127	27	\mathcal	\mathcal	PROPN
bracis-33593	127	28	{	{	PUNCT
bracis-33593	127	29	n}(s+a\tau	n}(s+a\tau	PROPN
bracis-33593	127	30	_	_	PUNCT
bracis-33593	128	1	i	i	PRON
bracis-33593	128	2	,	,	PUNCT
bracis-33593	128	3	\sigma	\sigma	PROPN
bracis-33593	128	4	_	_	PUNCT
bracis-33593	128	5	i)\	i)\	NOUN
bracis-33593	128	6	)	)	PUNCT
bracis-33593	128	7	with	with	ADP
bracis-33593	128	8	\(\tau	\(\tau	PROPN
bracis-33593	128	9	_	_	X
bracis-33593	128	10	i\	i\	NOUN
bracis-33593	128	11	)	)	PUNCT
bracis-33593	128	12	and	and	CCONJ
bracis-33593	128	13	\(\sigma	\(\sigma	NOUN
bracis-33593	128	14	_	_	PRON
bracis-33593	128	15	i\	i\	NOUN
bracis-33593	128	16	)	)	PUNCT
bracis-33593	128	17	being	be	AUX
bracis-33593	128	18	parameters	parameter	NOUN
bracis-33593	128	19	\(\theta	\(\theta	VERB
bracis-33593	128	20	_	_	PUNCT
bracis-33593	128	21	i\	i\	NOUN
bracis-33593	128	22	)	)	PUNCT
bracis-33593	128	23	of	of	ADP
bracis-33593	128	24	region	region	NOUN
bracis-33593	128	25	\(m_i\	\(m_i\	NOUN
bracis-33593	128	26	)	)	PUNCT
bracis-33593	128	27	where	where	SCONJ
bracis-33593	128	28	\(m_i	\(m_i	VERB
bracis-33593	128	29	\subset	\subset	PROPN
bracis-33593	129	1	\mathcal	\mathcal	PROPN
bracis-33593	129	2	{	{	PUNCT
bracis-33593	129	3	s	s	PROPN
bracis-33593	129	4	}	}	PUNCT
bracis-33593	129	5	;	;	PUNCT
bracis-33593	129	6	s	s	X
bracis-33593	129	7	\in	\in	ADJ
bracis-33593	129	8	m_i\	m_i\	NOUN
bracis-33593	129	9	)	)	PUNCT
bracis-33593	129	10	 	 	SPACE
bracis-33593	129	11	and	and	CCONJ
bracis-33593	129	12	\(s	\(s	PROPN
bracis-33593	129	13	'	'	PART
bracis-33593	129	14	\in	\in	PROPN
bracis-33593	129	15	\mathcal	\mathcal	PROPN
bracis-33593	129	16	{	{	PUNCT
bracis-33593	129	17	s}\	s}\	NUM
bracis-33593	129	18	)	)	PUNCT
bracis-33593	129	19	;	;	PUNCT
bracis-33593	129	20	\(a	\(a	PROPN
bracis-33593	129	21	\in	\in	PROPN
bracis-33593	129	22	\mathcal	\mathcal	PROPN
bracis-33593	129	23	{	{	PUNCT
bracis-33593	129	24	a}\	a}\	NUM
bracis-33593	129	25	)	)	PUNCT
bracis-33593	129	26	;	;	PUNCT
bracis-33593	129	27	this	this	PRON
bracis-33593	129	28	give	give	VERB
bracis-33593	129	29	us	we	PRON
bracis-33593	129	30	a	a	DET
bracis-33593	129	31	transition	transition	NOUN
bracis-33593	129	32	function	function	NOUN
bracis-33593	129	33	that	that	PRON
bracis-33593	129	34	is	be	AUX
bracis-33593	129	35	derived	derive	VERB
bracis-33593	129	36	from	from	ADP
bracis-33593	129	37	the	the	DET
bracis-33593	129	38	probability	probability	NOUN
bracis-33593	129	39	density	density	NOUN
bracis-33593	129	40	function	function	VERB
bracis-33593	129	41	dependent	dependent	ADJ
bracis-33593	129	42	of	of	ADP
bracis-33593	129	43	the	the	DET
bracis-33593	129	44	region	region	NOUN
bracis-33593	130	1	i	i	X
bracis-33593	130	2	:	:	PUNCT
bracis-33593	130	3	$	$	SYM
bracis-33593	130	4	$	$	SYM
bracis-33593	130	5	\begin{aligned	\begin{aligne	VERB
bracis-33593	130	6	}	}	PUNCT
bracis-33593	130	7	\mathcal	\mathcal	ADJ
bracis-33593	130	8	{	{	PUNCT
bracis-33593	130	9	p}_{\theta	p}_{\theta	NOUN
bracis-33593	130	10	}	}	PUNCT
bracis-33593	130	11	(	(	PUNCT
bracis-33593	130	12	s'|	s'|	NOUN
bracis-33593	130	13	s	s	PART
bracis-33593	130	14	,	,	PUNCT
bracis-33593	130	15	a	a	PRON
bracis-33593	130	16	)	)	PUNCT
bracis-33593	130	17	=	=	SYM
bracis-33593	130	18	\frac{1}{\sigma	\frac{1}{\sigma	PROPN
bracis-33593	130	19	\sqrt{2\pi	\sqrt{2\pi	NUM
bracis-33593	130	20	}	}	PUNCT
bracis-33593	130	21	}	}	PUNCT
bracis-33593	130	22	e^{-\frac{1}{2}(\frac{s'-(s+a\tau	e^{-\frac{1}{2}(\frac{s'-(s+a\tau	PROPN
bracis-33593	130	23	)	)	PUNCT
bracis-33593	130	24	}	}	PUNCT
bracis-33593	130	25	{	{	PUNCT
bracis-33593	130	26	\sigma	\sigma	PROPN
bracis-33593	130	27	}	}	PUNCT
bracis-33593	130	28	)	)	PUNCT
bracis-33593	130	29	^2	^2	ADP
bracis-33593	130	30	}	}	PUNCT
bracis-33593	130	31	\end{aligned}$$	\end{aligned}$$	X
bracis-33593	130	32	(	(	PUNCT
bracis-33593	130	33	4	4	NUM
bracis-33593	130	34	)	)	PUNCT
bracis-33593	130	35	the	the	DET
bracis-33593	130	36	probability	probability	NOUN
bracis-33593	130	37	defined	define	VERB
bracis-33593	130	38	by	by	ADP
bracis-33593	130	39	eq	eq	AUX
bracis-33593	130	40	.	.	AUX
bracis-33593	130	41	 	 	SPACE
bracis-33593	130	42	4	4	NUM
bracis-33593	130	43	will	will	AUX
bracis-33593	130	44	be	be	AUX
bracis-33593	130	45	used	use	VERB
bracis-33593	130	46	to	to	PART
bracis-33593	130	47	determine	determine	VERB
bracis-33593	130	48	the	the	DET
bracis-33593	130	49	likelihood	likelihood	NOUN
bracis-33593	130	50	function	function	NOUN
bracis-33593	130	51	,	,	PUNCT
bracis-33593	130	52	as	as	SCONJ
bracis-33593	130	53	elaborated	elaborate	VERB
bracis-33593	130	54	in	in	ADP
bracis-33593	130	55	sect	sect	NOUN
bracis-33593	130	56	.	.	PUNCT
bracis-33593	130	57	 	 	SPACE
bracis-33593	131	1	4	4	NUM
bracis-33593	131	2	.	.	PUNCT
bracis-33593	131	3	an	an	DET
bracis-33593	131	4	illustration	illustration	NOUN
bracis-33593	131	5	of	of	ADP
bracis-33593	131	6	such	such	DET
bracis-33593	131	7	an	an	DET
bracis-33593	131	8	environment	environment	NOUN
bracis-33593	131	9	can	can	AUX
bracis-33593	131	10	be	be	AUX
bracis-33593	131	11	observed	observe	VERB
bracis-33593	131	12	in	in	ADP
bracis-33593	131	13	fig	fig	NOUN
bracis-33593	131	14	.	.	PUNCT
bracis-33593	131	15	 	 	SPACE
bracis-33593	132	1	2	2	NUM
bracis-33593	132	2	,	,	PUNCT
bracis-33593	132	3	providing	provide	VERB
bracis-33593	132	4	a	a	DET
bracis-33593	132	5	visual	visual	ADJ
bracis-33593	132	6	representation	representation	NOUN
bracis-33593	132	7	of	of	ADP
bracis-33593	132	8	the	the	DET
bracis-33593	132	9	two	two	NUM
bracis-33593	132	10	-	-	PUNCT
bracis-33593	132	11	dimensional	dimensional	ADJ
bracis-33593	132	12	space	space	NOUN
bracis-33593	132	13	with	with	ADP
bracis-33593	132	14	segmented	segment	VERB
bracis-33593	132	15	dynamics	dynamic	NOUN
bracis-33593	132	16	.	.	PUNCT
bracis-33593	133	1	this	this	DET
bracis-33593	133	2	figure	figure	NOUN
bracis-33593	133	3	effectively	effectively	ADV
bracis-33593	133	4	encapsulates	encapsulate	VERB
bracis-33593	133	5	the	the	DET
bracis-33593	133	6	concept	concept	NOUN
bracis-33593	133	7	of	of	ADP
bracis-33593	133	8	distinct	distinct	ADJ
bracis-33593	133	9	regions	region	NOUN
bracis-33593	133	10	within	within	ADP
bracis-33593	133	11	the	the	DET
bracis-33593	133	12	state	state	NOUN
bracis-33593	133	13	space	space	NOUN
bracis-33593	133	14	,	,	PUNCT
bracis-33593	133	15	each	each	PRON
bracis-33593	133	16	characterized	characterize	VERB
bracis-33593	133	17	by	by	ADP
bracis-33593	133	18	unique	unique	ADJ
bracis-33593	133	19	parameters	parameter	NOUN
bracis-33593	133	20	of	of	ADP
bracis-33593	133	21	the	the	DET
bracis-33593	133	22	gaussian	gaussian	ADJ
bracis-33593	133	23	distribution	distribution	NOUN
bracis-33593	133	24	governing	govern	VERB
bracis-33593	133	25	uncertainty	uncertainty	NOUN
bracis-33593	133	26	in	in	ADP
bracis-33593	133	27	transitions	transition	NOUN
bracis-33593	133	28	.	.	PUNCT
bracis-33593	134	1	fig	fig	NOUN
bracis-33593	134	2	.	.	PUNCT
bracis-33593	135	1	2	2	X
bracis-33593	135	2	.	.	X
bracis-33593	135	3	a	a	DET
bracis-33593	135	4	representation	representation	NOUN
bracis-33593	135	5	of	of	ADP
bracis-33593	135	6	a	a	DET
bracis-33593	135	7	instance	instance	NOUN
bracis-33593	135	8	of	of	ADP
bracis-33593	135	9	environment	environment	NOUN
bracis-33593	135	10	,	,	PUNCT
bracis-33593	135	11	with	with	ADP
bracis-33593	135	12	the	the	DET
bracis-33593	135	13	goal	goal	NOUN
bracis-33593	135	14	states	state	NOUN
bracis-33593	135	15	represented	represent	VERB
bracis-33593	135	16	by	by	ADP
bracis-33593	135	17	the	the	DET
bracis-33593	135	18	square	square	NOUN
bracis-33593	135	19	with	with	ADP
bracis-33593	135	20	g	g	PROPN
bracis-33593	135	21	,	,	PUNCT
bracis-33593	135	22	and	and	CCONJ
bracis-33593	135	23	a	a	DET
bracis-33593	135	24	set	set	NOUN
bracis-33593	135	25	of	of	ADP
bracis-33593	135	26	initial	initial	ADJ
bracis-33593	135	27	states	state	NOUN
bracis-33593	135	28	represented	represent	VERB
bracis-33593	135	29	by	by	ADP
bracis-33593	135	30	the	the	DET
bracis-33593	135	31	white	white	ADJ
bracis-33593	135	32	rectangle	rectangle	NOUN
bracis-33593	135	33	,	,	PUNCT
bracis-33593	135	34	and	and	CCONJ
bracis-33593	135	35	the	the	DET
bracis-33593	135	36	two	two	NUM
bracis-33593	135	37	regions	region	NOUN
bracis-33593	135	38	of	of	ADP
bracis-33593	135	39	different	different	ADJ
bracis-33593	135	40	dynamics	dynamic	NOUN
bracis-33593	135	41	represented	represent	VERB
bracis-33593	135	42	with	with	ADP
bracis-33593	135	43	the	the	DET
bracis-33593	135	44	colors	color	NOUN
bracis-33593	135	45	yellow	yellow	ADJ
bracis-33593	135	46	and	and	CCONJ
bracis-33593	135	47	blue	blue	ADJ
bracis-33593	135	48	(	(	PUNCT
bracis-33593	135	49	color	color	NOUN
bracis-33593	135	50	figure	figure	NOUN
bracis-33593	135	51	online	online	ADV
bracis-33593	135	52	)	)	PUNCT
bracis-33593	135	53	full	full	ADJ
bracis-33593	135	54	size	size	NOUN
bracis-33593	135	55	image	image	NOUN
bracis-33593	135	56	by	by	ADP
bracis-33593	135	57	implementing	implement	VERB
bracis-33593	135	58	this	this	DET
bracis-33593	135	59	approach	approach	NOUN
bracis-33593	135	60	,	,	PUNCT
bracis-33593	135	61	we	we	PRON
bracis-33593	135	62	effectively	effectively	ADV
bracis-33593	135	63	model	model	VERB
bracis-33593	135	64	environments	environment	NOUN
bracis-33593	135	65	where	where	SCONJ
bracis-33593	135	66	dynamics	dynamic	NOUN
bracis-33593	135	67	vary	vary	VERB
bracis-33593	135	68	between	between	ADP
bracis-33593	135	69	different	different	ADJ
bracis-33593	135	70	regions	region	NOUN
bracis-33593	135	71	.	.	PUNCT
bracis-33593	136	1	adjusting	adjust	VERB
bracis-33593	136	2	the	the	DET
bracis-33593	136	3	parameters	parameter	NOUN
bracis-33593	136	4	of	of	ADP
bracis-33593	136	5	the	the	DET
bracis-33593	136	6	gaussian	gaussian	ADJ
bracis-33593	136	7	distribution	distribution	NOUN
bracis-33593	136	8	enables	enable	VERB
bracis-33593	136	9	us	we	PRON
bracis-33593	136	10	to	to	PART
bracis-33593	136	11	replicate	replicate	VERB
bracis-33593	136	12	diverse	diverse	ADJ
bracis-33593	136	13	levels	level	NOUN
bracis-33593	136	14	of	of	ADP
bracis-33593	136	15	uncertainty	uncertainty	NOUN
bracis-33593	136	16	and	and	CCONJ
bracis-33593	136	17	variability	variability	NOUN
bracis-33593	136	18	within	within	ADP
bracis-33593	136	19	each	each	DET
bracis-33593	136	20	region	region	NOUN
bracis-33593	136	21	.	.	PUNCT
bracis-33593	137	1	consequently	consequently	ADV
bracis-33593	137	2	,	,	PUNCT
bracis-33593	137	3	agents	agent	NOUN
bracis-33593	137	4	must	must	AUX
bracis-33593	137	5	adapt	adapt	VERB
bracis-33593	137	6	their	their	PRON
bracis-33593	137	7	strategies	strategy	NOUN
bracis-33593	137	8	accordingly	accordingly	ADV
bracis-33593	137	9	as	as	SCONJ
bracis-33593	137	10	they	they	PRON
bracis-33593	137	11	navigate	navigate	VERB
bracis-33593	137	12	through	through	ADP
bracis-33593	137	13	the	the	DET
bracis-33593	137	14	environment	environment	NOUN
bracis-33593	137	15	.	.	PUNCT
bracis-33593	138	1	this	this	DET
bracis-33593	138	2	simulation	simulation	NOUN
bracis-33593	138	3	framework	framework	NOUN
bracis-33593	138	4	offers	offer	VERB
bracis-33593	138	5	a	a	DET
bracis-33593	138	6	versatile	versatile	ADJ
bracis-33593	138	7	platform	platform	NOUN
bracis-33593	138	8	for	for	ADP
bracis-33593	138	9	investigating	investigate	VERB
bracis-33593	138	10	and	and	CCONJ
bracis-33593	138	11	evaluating	evaluate	VERB
bracis-33593	138	12	model	model	NOUN
bracis-33593	138	13	-	-	PUNCT
bracis-33593	138	14	based	base	VERB
bracis-33593	138	15	reinforcement	reinforcement	NOUN
bracis-33593	138	16	learning	learning	NOUN
bracis-33593	138	17	algorithms	algorithm	NOUN
bracis-33593	138	18	in	in	ADP
bracis-33593	138	19	scenarios	scenario	NOUN
bracis-33593	138	20	where	where	SCONJ
bracis-33593	138	21	the	the	DET
bracis-33593	138	22	state	state	NOUN
bracis-33593	138	23	space	space	NOUN
bracis-33593	138	24	exhibits	exhibit	VERB
bracis-33593	138	25	heterogeneous	heterogeneous	ADJ
bracis-33593	138	26	dynamics	dynamic	NOUN
bracis-33593	138	27	.	.	PUNCT
bracis-33593	139	1	through	through	ADP
bracis-33593	139	2	experimentation	experimentation	NOUN
bracis-33593	139	3	and	and	CCONJ
bracis-33593	139	4	analysis	analysis	NOUN
bracis-33593	139	5	within	within	ADP
bracis-33593	139	6	this	this	DET
bracis-33593	139	7	environment	environment	NOUN
bracis-33593	139	8	,	,	PUNCT
bracis-33593	139	9	researchers	researcher	NOUN
bracis-33593	139	10	gain	gain	VERB
bracis-33593	139	11	valuable	valuable	ADJ
bracis-33593	139	12	insights	insight	NOUN
bracis-33593	139	13	into	into	ADP
bracis-33593	139	14	the	the	DET
bracis-33593	139	15	effectiveness	effectiveness	NOUN
bracis-33593	139	16	and	and	CCONJ
bracis-33593	139	17	adaptability	adaptability	NOUN
bracis-33593	139	18	of	of	ADP
bracis-33593	139	19	different	different	ADJ
bracis-33593	139	20	learning	learning	NOUN
bracis-33593	139	21	approaches	approach	NOUN
bracis-33593	139	22	in	in	ADP
bracis-33593	139	23	complex	complex	ADJ
bracis-33593	139	24	real	real	ADJ
bracis-33593	139	25	-	-	PUNCT
bracis-33593	139	26	world	world	NOUN
bracis-33593	139	27	settings	setting	NOUN
bracis-33593	139	28	.	.	PUNCT
bracis-33593	140	1	5.1	5.1	NUM
bracis-33593	140	2	functions	function	NOUN
bracis-33593	140	3	aproximators	aproximator	NOUN
bracis-33593	140	4	in	in	ADP
bracis-33593	140	5	our	our	PRON
bracis-33593	140	6	approach	approach	NOUN
bracis-33593	140	7	,	,	PUNCT
bracis-33593	140	8	we	we	PRON
bracis-33593	140	9	leverage	leverage	VERB
bracis-33593	140	10	the	the	DET
bracis-33593	140	11	likelihood	likelihood	NOUN
bracis-33593	140	12	estimators	estimator	NOUN
bracis-33593	140	13	to	to	PART
bracis-33593	140	14	optimize	optimize	VERB
bracis-33593	140	15	both	both	DET
bracis-33593	140	16	the	the	DET
bracis-33593	140	17	parameters	parameter	NOUN
bracis-33593	140	18	of	of	ADP
bracis-33593	140	19	the	the	DET
bracis-33593	140	20	\(\beta	\(\beta	PROPN
bracis-33593	140	21	\	\	PROPN
bracis-33593	140	22	)	)	PUNCT
bracis-33593	140	23	function	function	NOUN
bracis-33593	140	24	and	and	CCONJ
bracis-33593	140	25	the	the	DET
bracis-33593	140	26	k	k	NOUN
bracis-33593	140	27	-	-	NOUN
bracis-33593	140	28	models	model	NOUN
bracis-33593	140	29	,	,	PUNCT
bracis-33593	140	30	which	which	PRON
bracis-33593	140	31	can	can	AUX
bracis-33593	140	32	also	also	ADV
bracis-33593	140	33	be	be	AUX
bracis-33593	140	34	represented	represent	VERB
bracis-33593	140	35	by	by	ADP
bracis-33593	140	36	parameterized	parameterized	ADJ
bracis-33593	140	37	functions	function	NOUN
bracis-33593	140	38	.	.	PUNCT
bracis-33593	141	1	for	for	ADP
bracis-33593	141	2	simplicity	simplicity	NOUN
bracis-33593	141	3	and	and	CCONJ
bracis-33593	141	4	flexibility	flexibility	NOUN
bracis-33593	141	5	,	,	PUNCT
bracis-33593	141	6	in	in	ADP
bracis-33593	141	7	our	our	PRON
bracis-33593	141	8	experiments	experiment	NOUN
bracis-33593	141	9	,	,	PUNCT
bracis-33593	141	10	we	we	PRON
bracis-33593	141	11	employed	employ	VERB
bracis-33593	141	12	a	a	DET
bracis-33593	141	13	feedforward	feedforward	ADJ
bracis-33593	141	14	neural	neural	ADJ
bracis-33593	141	15	network	network	NOUN
bracis-33593	141	16	with	with	ADP
bracis-33593	141	17	an	an	DET
bracis-33593	141	18	softmax	softmax	NOUN
bracis-33593	141	19	activation	activation	NOUN
bracis-33593	141	20	function	function	NOUN
bracis-33593	141	21	in	in	ADP
bracis-33593	141	22	the	the	DET
bracis-33593	141	23	output	output	NOUN
bracis-33593	141	24	layer	layer	NOUN
bracis-33593	141	25	to	to	PART
bracis-33593	141	26	represent	represent	VERB
bracis-33593	141	27	the	the	DET
bracis-33593	141	28	\(\beta	\(\beta	PROPN
bracis-33593	141	29	\	\	NOUN
bracis-33593	141	30	)	)	PUNCT
bracis-33593	141	31	functions	function	NOUN
bracis-33593	141	32	.	.	PUNCT
bracis-33593	142	1	additionally	additionally	ADV
bracis-33593	142	2	,	,	PUNCT
bracis-33593	142	3	we	we	PRON
bracis-33593	142	4	assumed	assume	VERB
bracis-33593	142	5	a	a	DET
bracis-33593	142	6	known	know	VERB
bracis-33593	142	7	distribution	distribution	NOUN
bracis-33593	142	8	function	function	NOUN
bracis-33593	142	9	for	for	ADP
bracis-33593	142	10	the	the	DET
bracis-33593	142	11	k	k	NOUN
bracis-33593	142	12	-	-	NOUN
bracis-33593	142	13	models	model	NOUN
bracis-33593	142	14	,	,	PUNCT
bracis-33593	142	15	specifically	specifically	ADV
bracis-33593	142	16	a	a	DET
bracis-33593	142	17	gaussian	gaussian	ADJ
bracis-33593	142	18	distribution	distribution	NOUN
bracis-33593	142	19	.	.	PUNCT
bracis-33593	143	1	the	the	DET
bracis-33593	143	2	parameters	parameter	NOUN
bracis-33593	143	3	of	of	ADP
bracis-33593	143	4	the	the	DET
bracis-33593	143	5	neural	neural	ADJ
bracis-33593	143	6	network	network	NOUN
bracis-33593	143	7	,	,	PUNCT
bracis-33593	143	8	along	along	ADP
bracis-33593	143	9	with	with	ADP
bracis-33593	143	10	the	the	DET
bracis-33593	143	11	means	mean	NOUN
bracis-33593	143	12	and	and	CCONJ
bracis-33593	143	13	variances	variance	NOUN
bracis-33593	143	14	of	of	ADP
bracis-33593	143	15	the	the	DET
bracis-33593	143	16	gaussian	gaussian	ADJ
bracis-33593	143	17	distributions	distribution	NOUN
bracis-33593	143	18	representing	represent	VERB
bracis-33593	143	19	the	the	DET
bracis-33593	143	20	k	k	NOUN
bracis-33593	143	21	-	-	NOUN
bracis-33593	143	22	models	model	NOUN
bracis-33593	143	23	,	,	PUNCT
bracis-33593	143	24	collectively	collectively	ADV
bracis-33593	143	25	form	form	VERB
bracis-33593	143	26	the	the	DET
bracis-33593	143	27	parameter	parameter	NOUN
bracis-33593	143	28	set	set	VERB
bracis-33593	143	29	\(\vartheta	\(\vartheta	NOUN
bracis-33593	143	30	\	\	PROPN
bracis-33593	143	31	)	)	PUNCT
bracis-33593	143	32	that	that	SCONJ
bracis-33593	143	33	we	we	PRON
bracis-33593	143	34	aim	aim	VERB
bracis-33593	143	35	to	to	PART
bracis-33593	143	36	estimate	estimate	VERB
bracis-33593	143	37	using	use	VERB
bracis-33593	143	38	the	the	DET
bracis-33593	143	39	likelihood	likelihood	NOUN
bracis-33593	143	40	-	-	PUNCT
bracis-33593	143	41	based	base	VERB
bracis-33593	143	42	optimization	optimization	NOUN
bracis-33593	143	43	method	method	NOUN
bracis-33593	143	44	described	describe	VERB
bracis-33593	143	45	previously	previously	ADV
bracis-33593	143	46	.	.	PUNCT
bracis-33593	144	1	replacing	replace	VERB
bracis-33593	144	2	the	the	DET
bracis-33593	144	3	structural	structural	ADJ
bracis-33593	144	4	assumptions	assumption	NOUN
bracis-33593	144	5	for	for	ADP
bracis-33593	144	6	the	the	DET
bracis-33593	144	7	k	k	PROPN
bracis-33593	144	8	-	-	PUNCT
bracis-33593	144	9	model	model	NOUN
bracis-33593	144	10	’s	’s	PART
bracis-33593	144	11	functions	function	NOUN
bracis-33593	144	12	with	with	ADP
bracis-33593	144	13	a	a	DET
bracis-33593	144	14	universal	universal	ADJ
bracis-33593	144	15	function	function	NOUN
bracis-33593	144	16	approximation	approximation	NOUN
bracis-33593	144	17	,	,	PUNCT
bracis-33593	144	18	similar	similar	ADJ
bracis-33593	144	19	to	to	ADP
bracis-33593	144	20	what	what	PRON
bracis-33593	144	21	we	we	PRON
bracis-33593	144	22	’ve	’ve	AUX
bracis-33593	144	23	done	do	VERB
bracis-33593	144	24	for	for	ADP
bracis-33593	144	25	the	the	DET
bracis-33593	144	26	\(\beta	\(\beta	PROPN
bracis-33593	144	27	\	\	ADJ
bracis-33593	144	28	)	)	PUNCT
bracis-33593	144	29	function	function	NOUN
bracis-33593	144	30	,	,	PUNCT
bracis-33593	144	31	is	be	AUX
bracis-33593	144	32	a	a	DET
bracis-33593	144	33	natural	natural	ADJ
bracis-33593	144	34	step	step	NOUN
bracis-33593	144	35	towards	towards	ADP
bracis-33593	144	36	generalizing	generalize	VERB
bracis-33593	144	37	this	this	DET
bracis-33593	144	38	solution	solution	NOUN
bracis-33593	144	39	for	for	ADP
bracis-33593	144	40	more	more	ADV
bracis-33593	144	41	practical	practical	ADJ
bracis-33593	144	42	problems	problem	NOUN
bracis-33593	144	43	,	,	PUNCT
bracis-33593	144	44	and	and	CCONJ
bracis-33593	144	45	for	for	ADP
bracis-33593	144	46	that	that	PRON
bracis-33593	144	47	,	,	PUNCT
bracis-33593	144	48	no	no	DET
bracis-33593	144	49	changes	change	NOUN
bracis-33593	144	50	in	in	ADP
bracis-33593	144	51	the	the	DET
bracis-33593	144	52	likelihood	likelihood	NOUN
bracis-33593	144	53	estimator	estimator	NOUN
bracis-33593	144	54	is	be	AUX
bracis-33593	144	55	necessary	necessary	ADJ
bracis-33593	144	56	,	,	PUNCT
bracis-33593	144	57	since	since	SCONJ
bracis-33593	144	58	it	it	PRON
bracis-33593	144	59	’s	’	VERB
bracis-33593	144	60	agnostic	agnostic	ADJ
bracis-33593	144	61	to	to	ADP
bracis-33593	144	62	the	the	DET
bracis-33593	144	63	functions	function	NOUN
bracis-33593	144	64	structure	structure	NOUN
bracis-33593	144	65	,	,	PUNCT
bracis-33593	144	66	optimizing	optimize	VERB
bracis-33593	144	67	the	the	DET
bracis-33593	144	68	parameters	parameter	NOUN
bracis-33593	144	69	directly	directly	ADV
bracis-33593	144	70	.	.	PUNCT
bracis-33593	145	1	fig	fig	NOUN
bracis-33593	145	2	.	.	PUNCT
bracis-33593	146	1	3	3	X
bracis-33593	146	2	.	.	X
bracis-33593	146	3	a	a	DET
bracis-33593	146	4	representation	representation	NOUN
bracis-33593	146	5	of	of	ADP
bracis-33593	146	6	an	an	DET
bracis-33593	146	7	episode	episode	NOUN
bracis-33593	146	8	,	,	PUNCT
bracis-33593	146	9	showing	show	VERB
bracis-33593	146	10	the	the	DET
bracis-33593	146	11	steps	step	NOUN
bracis-33593	146	12	where	where	SCONJ
bracis-33593	146	13	the	the	DET
bracis-33593	146	14	model	model	NOUN
bracis-33593	146	15	is	be	AUX
bracis-33593	146	16	updated	update	VERB
bracis-33593	146	17	and	and	CCONJ
bracis-33593	146	18	a	a	DET
bracis-33593	146	19	new	new	ADJ
bracis-33593	146	20	plan	plan	NOUN
bracis-33593	146	21	is	be	AUX
bracis-33593	146	22	generated	generate	VERB
bracis-33593	146	23	full	full	ADJ
bracis-33593	146	24	size	size	NOUN
bracis-33593	146	25	image	image	NOUN
bracis-33593	146	26	the	the	DET
bracis-33593	146	27	dynamics	dynamic	NOUN
bracis-33593	146	28	for	for	ADP
bracis-33593	146	29	updating	update	VERB
bracis-33593	146	30	the	the	DET
bracis-33593	146	31	models	model	NOUN
bracis-33593	146	32	based	base	VERB
bracis-33593	146	33	on	on	ADP
bracis-33593	146	34	new	new	ADJ
bracis-33593	146	35	experience	experience	NOUN
bracis-33593	146	36	and	and	CCONJ
bracis-33593	146	37	the	the	DET
bracis-33593	146	38	generation	generation	NOUN
bracis-33593	146	39	of	of	ADP
bracis-33593	146	40	a	a	DET
bracis-33593	146	41	new	new	ADJ
bracis-33593	146	42	plan	plan	NOUN
bracis-33593	146	43	base	base	NOUN
bracis-33593	146	44	on	on	ADP
bracis-33593	146	45	the	the	DET
bracis-33593	146	46	new	new	ADJ
bracis-33593	146	47	model	model	NOUN
bracis-33593	146	48	can	can	AUX
bracis-33593	146	49	be	be	AUX
bracis-33593	146	50	visualized	visualize	VERB
bracis-33593	146	51	on	on	ADP
bracis-33593	146	52	fig	fig	NOUN
bracis-33593	146	53	.	.	PUNCT
bracis-33593	146	54	 	 	SPACE
bracis-33593	147	1	3	3	NUM
bracis-33593	147	2	that	that	PRON
bracis-33593	147	3	show	show	VERB
bracis-33593	147	4	how	how	SCONJ
bracis-33593	147	5	an	an	DET
bracis-33593	147	6	episode	episode	NOUN
bracis-33593	147	7	starts	start	VERB
bracis-33593	147	8	on	on	ADP
bracis-33593	147	9	\(s_0\	\(s_0\	PROPN
bracis-33593	147	10	)	)	PUNCT
bracis-33593	147	11	and	and	CCONJ
bracis-33593	147	12	a	a	DET
bracis-33593	147	13	plan	plan	NOUN
bracis-33593	147	14	is	be	AUX
bracis-33593	147	15	generated	generate	VERB
bracis-33593	147	16	to	to	PART
bracis-33593	147	17	start	start	VERB
bracis-33593	147	18	guiding	guide	VERB
bracis-33593	147	19	the	the	DET
bracis-33593	147	20	agent	agent	NOUN
bracis-33593	147	21	’s	’s	PART
bracis-33593	147	22	initial	initial	ADJ
bracis-33593	147	23	actions	action	NOUN
bracis-33593	147	24	,	,	PUNCT
bracis-33593	147	25	and	and	CCONJ
bracis-33593	147	26	after	after	ADP
bracis-33593	147	27	a	a	DET
bracis-33593	147	28	certain	certain	ADJ
bracis-33593	147	29	number	number	NOUN
bracis-33593	147	30	of	of	ADP
bracis-33593	147	31	steps	step	NOUN
bracis-33593	147	32	,	,	PUNCT
bracis-33593	147	33	the	the	DET
bracis-33593	147	34	model	model	NOUN
bracis-33593	147	35	can	can	AUX
bracis-33593	147	36	be	be	AUX
bracis-33593	147	37	updated	update	VERB
bracis-33593	147	38	based	base	VERB
bracis-33593	147	39	on	on	ADP
bracis-33593	147	40	the	the	DET
bracis-33593	147	41	new	new	ADJ
bracis-33593	147	42	gained	gain	VERB
bracis-33593	147	43	information	information	NOUN
bracis-33593	147	44	,	,	PUNCT
bracis-33593	147	45	which	which	PRON
bracis-33593	147	46	enable	enable	VERB
bracis-33593	147	47	the	the	DET
bracis-33593	147	48	agent	agent	NOUN
bracis-33593	147	49	to	to	PART
bracis-33593	147	50	come	come	VERB
bracis-33593	147	51	with	with	ADP
bracis-33593	147	52	a	a	DET
bracis-33593	147	53	new	new	ADJ
bracis-33593	147	54	plan	plan	NOUN
bracis-33593	147	55	,	,	PUNCT
bracis-33593	147	56	and	and	CCONJ
bracis-33593	147	57	this	this	PRON
bracis-33593	147	58	goes	go	VERB
bracis-33593	147	59	on	on	ADP
bracis-33593	147	60	,	,	PUNCT
bracis-33593	147	61	until	until	SCONJ
bracis-33593	147	62	the	the	DET
bracis-33593	147	63	agent	agent	NOUN
bracis-33593	147	64	reaches	reach	VERB
bracis-33593	147	65	the	the	DET
bracis-33593	147	66	goal	goal	NOUN
bracis-33593	147	67	.	.	PUNCT
bracis-33593	148	1	there	there	PRON
bracis-33593	148	2	are	be	VERB
bracis-33593	148	3	two	two	NUM
bracis-33593	148	4	aspects	aspect	NOUN
bracis-33593	148	5	that	that	PRON
bracis-33593	148	6	can	can	AUX
bracis-33593	148	7	be	be	AUX
bracis-33593	148	8	adjusted	adjust	VERB
bracis-33593	148	9	in	in	ADP
bracis-33593	148	10	our	our	PRON
bracis-33593	148	11	approach	approach	NOUN
bracis-33593	148	12	that	that	PRON
bracis-33593	148	13	can	can	AUX
bracis-33593	148	14	change	change	VERB
bracis-33593	148	15	the	the	DET
bracis-33593	148	16	impact	impact	NOUN
bracis-33593	148	17	of	of	ADP
bracis-33593	148	18	the	the	DET
bracis-33593	148	19	learning	learning	NOUN
bracis-33593	148	20	of	of	ADP
bracis-33593	148	21	the	the	DET
bracis-33593	148	22	models	model	NOUN
bracis-33593	148	23	.	.	PUNCT
bracis-33593	149	1	the	the	DET
bracis-33593	149	2	first	first	ADJ
bracis-33593	149	3	one	one	NOUN
bracis-33593	149	4	is	be	AUX
bracis-33593	149	5	the	the	DET
bracis-33593	149	6	memory	memory	NOUN
bracis-33593	149	7	size	size	NOUN
bracis-33593	149	8	,	,	PUNCT
bracis-33593	149	9	since	since	SCONJ
bracis-33593	149	10	we	we	PRON
bracis-33593	149	11	use	use	VERB
bracis-33593	149	12	real	real	ADJ
bracis-33593	149	13	past	past	ADJ
bracis-33593	149	14	experience	experience	NOUN
bracis-33593	149	15	to	to	PART
bracis-33593	149	16	update	update	VERB
bracis-33593	149	17	the	the	DET
bracis-33593	149	18	parameters	parameter	NOUN
bracis-33593	149	19	via	via	ADP
bracis-33593	149	20	likelihood	likelihood	NOUN
bracis-33593	149	21	,	,	PUNCT
bracis-33593	149	22	we	we	PRON
bracis-33593	149	23	have	have	AUX
bracis-33593	149	24	decided	decide	VERB
bracis-33593	149	25	to	to	PART
bracis-33593	149	26	specify	specify	VERB
bracis-33593	149	27	the	the	DET
bracis-33593	149	28	size	size	NOUN
bracis-33593	149	29	of	of	ADP
bracis-33593	149	30	the	the	DET
bracis-33593	149	31	memory	memory	NOUN
bracis-33593	149	32	,	,	PUNCT
bracis-33593	149	33	in	in	ADP
bracis-33593	149	34	other	other	ADJ
bracis-33593	149	35	words	word	NOUN
bracis-33593	149	36	,	,	PUNCT
bracis-33593	149	37	how	how	SCONJ
bracis-33593	149	38	many	many	ADJ
bracis-33593	149	39	experiences	experience	VERB
bracis-33593	149	40	the	the	DET
bracis-33593	149	41	agent	agent	NOUN
bracis-33593	149	42	will	will	AUX
bracis-33593	149	43	remember	remember	VERB
bracis-33593	149	44	,	,	PUNCT
bracis-33593	149	45	and	and	CCONJ
bracis-33593	149	46	use	use	VERB
bracis-33593	149	47	to	to	PART
bracis-33593	149	48	recalculate	recalculate	VERB
bracis-33593	149	49	the	the	DET
bracis-33593	149	50	likelihood	likelihood	NOUN
bracis-33593	149	51	.	.	PUNCT
bracis-33593	150	1	as	as	SCONJ
bracis-33593	150	2	the	the	DET
bracis-33593	150	3	size	size	NOUN
bracis-33593	150	4	increases	increase	VERB
bracis-33593	150	5	,	,	PUNCT
bracis-33593	150	6	the	the	PRON
bracis-33593	150	7	better	well	ADJ
bracis-33593	150	8	the	the	DET
bracis-33593	150	9	updates	update	NOUN
bracis-33593	150	10	reflects	reflect	VERB
bracis-33593	150	11	the	the	DET
bracis-33593	150	12	reality	reality	NOUN
bracis-33593	150	13	,	,	PUNCT
bracis-33593	150	14	but	but	CCONJ
bracis-33593	150	15	given	give	VERB
bracis-33593	150	16	a	a	DET
bracis-33593	150	17	infinity	infinity	NOUN
bracis-33593	150	18	horizon	horizon	NOUN
bracis-33593	150	19	problem	problem	NOUN
bracis-33593	150	20	,	,	PUNCT
bracis-33593	150	21	this	this	PRON
bracis-33593	150	22	could	could	AUX
bracis-33593	150	23	became	become	VERB
bracis-33593	150	24	a	a	DET
bracis-33593	150	25	problem	problem	NOUN
bracis-33593	150	26	.	.	PUNCT
bracis-33593	151	1	the	the	DET
bracis-33593	151	2	second	second	ADJ
bracis-33593	151	3	relevant	relevant	ADJ
bracis-33593	151	4	aspect	aspect	NOUN
bracis-33593	151	5	is	be	AUX
bracis-33593	151	6	the	the	DET
bracis-33593	151	7	frequency	frequency	NOUN
bracis-33593	151	8	in	in	ADP
bracis-33593	151	9	which	which	PRON
bracis-33593	151	10	the	the	DET
bracis-33593	151	11	models	model	NOUN
bracis-33593	151	12	are	be	AUX
bracis-33593	151	13	updated	update	VERB
bracis-33593	151	14	.	.	PUNCT
bracis-33593	152	1	we	we	PRON
bracis-33593	152	2	control	control	VERB
bracis-33593	152	3	that	that	DET
bracis-33593	152	4	frequency	frequency	NOUN
bracis-33593	152	5	by	by	ADP
bracis-33593	152	6	specifying	specify	VERB
bracis-33593	152	7	the	the	DET
bracis-33593	152	8	number	number	NOUN
bracis-33593	152	9	of	of	ADP
bracis-33593	152	10	steps	step	NOUN
bracis-33593	152	11	that	that	PRON
bracis-33593	152	12	the	the	DET
bracis-33593	152	13	agent	agent	NOUN
bracis-33593	152	14	will	will	AUX
bracis-33593	152	15	be	be	AUX
bracis-33593	152	16	forced	force	VERB
bracis-33593	152	17	to	to	PART
bracis-33593	152	18	exploit	exploit	VERB
bracis-33593	152	19	the	the	DET
bracis-33593	152	20	models	model	NOUN
bracis-33593	152	21	after	after	ADP
bracis-33593	152	22	each	each	DET
bracis-33593	152	23	update	update	NOUN
bracis-33593	152	24	.	.	PUNCT
bracis-33593	153	1	so	so	ADV
bracis-33593	153	2	every	every	DET
bracis-33593	153	3	fixed	fix	VERB
bracis-33593	153	4	number	number	NOUN
bracis-33593	153	5	of	of	ADP
bracis-33593	153	6	steps	step	NOUN
bracis-33593	153	7	,	,	PUNCT
bracis-33593	153	8	the	the	DET
bracis-33593	153	9	experience	experience	NOUN
bracis-33593	153	10	collected	collect	VERB
bracis-33593	153	11	are	be	AUX
bracis-33593	153	12	added	add	VERB
bracis-33593	153	13	to	to	ADP
bracis-33593	153	14	the	the	DET
bracis-33593	153	15	memory	memory	NOUN
bracis-33593	153	16	and	and	CCONJ
bracis-33593	153	17	the	the	DET
bracis-33593	153	18	parameters	parameter	NOUN
bracis-33593	153	19	can	can	AUX
bracis-33593	153	20	be	be	AUX
bracis-33593	153	21	recalculated	recalculate	VERB
bracis-33593	153	22	afterwards	afterwards	ADV
bracis-33593	153	23	with	with	ADP
bracis-33593	153	24	the	the	DET
bracis-33593	153	25	extended	extended	ADJ
bracis-33593	153	26	trajectory	trajectory	NOUN
bracis-33593	153	27	.	.	PUNCT
bracis-33593	154	1	with	with	ADP
bracis-33593	154	2	all	all	PRON
bracis-33593	154	3	that	that	PRON
bracis-33593	154	4	put	put	VERB
bracis-33593	154	5	together	together	ADV
bracis-33593	154	6	we	we	PRON
bracis-33593	154	7	have	have	VERB
bracis-33593	154	8	a	a	DET
bracis-33593	154	9	solution	solution	NOUN
bracis-33593	154	10	that	that	PRON
bracis-33593	154	11	start	start	VERB
bracis-33593	154	12	roaming	roam	VERB
bracis-33593	154	13	for	for	ADP
bracis-33593	154	14	some	some	DET
bracis-33593	154	15	steps	step	NOUN
bracis-33593	154	16	and	and	CCONJ
bracis-33593	154	17	then	then	ADV
bracis-33593	154	18	uses	use	VERB
bracis-33593	154	19	its	its	PRON
bracis-33593	154	20	short	short	ADJ
bracis-33593	154	21	experience	experience	NOUN
bracis-33593	154	22	to	to	PART
bracis-33593	154	23	fit	fit	VERB
bracis-33593	154	24	a	a	DET
bracis-33593	154	25	model	model	NOUN
bracis-33593	154	26	,	,	PUNCT
bracis-33593	154	27	and	and	CCONJ
bracis-33593	154	28	then	then	ADV
bracis-33593	154	29	uses	use	VERB
bracis-33593	154	30	the	the	DET
bracis-33593	154	31	model	model	NOUN
bracis-33593	154	32	to	to	PART
bracis-33593	154	33	plan	plan	VERB
bracis-33593	154	34	the	the	DET
bracis-33593	154	35	next	next	ADJ
bracis-33593	154	36	actions	action	NOUN
bracis-33593	154	37	generating	generate	VERB
bracis-33593	154	38	new	new	ADJ
bracis-33593	154	39	experiences	experience	NOUN
bracis-33593	154	40	.	.	PUNCT
bracis-33593	155	1	following	follow	VERB
bracis-33593	155	2	this	this	DET
bracis-33593	155	3	loop	loop	NOUN
bracis-33593	155	4	for	for	ADP
bracis-33593	155	5	some	some	DET
bracis-33593	155	6	time	time	NOUN
bracis-33593	155	7	,	,	PUNCT
bracis-33593	155	8	eventually	eventually	ADV
bracis-33593	155	9	the	the	DET
bracis-33593	155	10	models	model	NOUN
bracis-33593	155	11	are	be	AUX
bracis-33593	155	12	able	able	ADJ
bracis-33593	155	13	to	to	PART
bracis-33593	155	14	get	get	VERB
bracis-33593	155	15	precise	precise	ADJ
bracis-33593	155	16	and	and	CCONJ
bracis-33593	155	17	the	the	DET
bracis-33593	155	18	plans	plan	NOUN
bracis-33593	155	19	actually	actually	ADV
bracis-33593	155	20	leads	lead	VERB
bracis-33593	155	21	to	to	ADP
bracis-33593	155	22	the	the	DET
bracis-33593	155	23	goal	goal	NOUN
bracis-33593	155	24	.	.	PUNCT
bracis-33593	156	1	fig	fig	NOUN
bracis-33593	156	2	.	.	PUNCT
bracis-33593	157	1	4	4	X
bracis-33593	157	2	.	.	X
bracis-33593	157	3	in	in	ADP
bracis-33593	157	4	blue	blue	ADJ
bracis-33593	157	5	is	be	AUX
bracis-33593	157	6	the	the	DET
bracis-33593	157	7	three	three	NUM
bracis-33593	157	8	step	step	NOUN
bracis-33593	157	9	tree	tree	NOUN
bracis-33593	157	10	plan	plan	NOUN
bracis-33593	157	11	generated	generate	VERB
bracis-33593	157	12	with	with	ADP
bracis-33593	157	13	the	the	DET
bracis-33593	157	14	models	model	NOUN
bracis-33593	157	15	learned	learn	VERB
bracis-33593	157	16	trough	trough	NOUN
bracis-33593	157	17	a	a	DET
bracis-33593	157	18	episode	episode	NOUN
bracis-33593	157	19	,	,	PUNCT
bracis-33593	157	20	on	on	ADP
bracis-33593	157	21	the	the	DET
bracis-33593	157	22	left	left	NOUN
bracis-33593	157	23	is	be	AUX
bracis-33593	157	24	the	the	DET
bracis-33593	157	25	tree	tree	NOUN
bracis-33593	157	26	generated	generate	VERB
bracis-33593	157	27	in	in	ADP
bracis-33593	157	28	the	the	DET
bracis-33593	157	29	beginning	beginning	NOUN
bracis-33593	157	30	and	and	CCONJ
bracis-33593	157	31	in	in	ADP
bracis-33593	157	32	the	the	DET
bracis-33593	157	33	right	right	NOUN
bracis-33593	157	34	at	at	ADP
bracis-33593	157	35	the	the	DET
bracis-33593	157	36	end	end	NOUN
bracis-33593	157	37	of	of	ADP
bracis-33593	157	38	the	the	DET
bracis-33593	157	39	episode	episode	NOUN
bracis-33593	157	40	.	.	PUNCT
bracis-33593	158	1	the	the	DET
bracis-33593	158	2	red	red	ADJ
bracis-33593	158	3	line	line	NOUN
bracis-33593	158	4	shows	show	VERB
bracis-33593	158	5	the	the	DET
bracis-33593	158	6	entire	entire	ADJ
bracis-33593	158	7	path	path	NOUN
bracis-33593	158	8	taken	take	VERB
bracis-33593	158	9	in	in	ADP
bracis-33593	158	10	the	the	DET
bracis-33593	158	11	episode	episode	NOUN
bracis-33593	158	12	.	.	PUNCT
bracis-33593	159	1	(	(	PUNCT
bracis-33593	159	2	color	color	NOUN
bracis-33593	159	3	figure	figure	NOUN
bracis-33593	159	4	online	online	ADV
bracis-33593	159	5	)	)	PUNCT
bracis-33593	159	6	full	full	ADJ
bracis-33593	159	7	size	size	NOUN
bracis-33593	159	8	image	image	NOUN
bracis-33593	159	9	the	the	DET
bracis-33593	159	10	fig	fig	NOUN
bracis-33593	159	11	.	.	PUNCT
bracis-33593	159	12	 	 	SPACE
bracis-33593	159	13	4	4	NUM
bracis-33593	159	14	show	show	VERB
bracis-33593	159	15	us	we	PRON
bracis-33593	159	16	two	two	NUM
bracis-33593	159	17	examples	example	NOUN
bracis-33593	159	18	of	of	ADP
bracis-33593	159	19	tree	tree	NOUN
bracis-33593	159	20	plan	plan	NOUN
bracis-33593	159	21	generated	generate	VERB
bracis-33593	159	22	with	with	ADP
bracis-33593	159	23	the	the	DET
bracis-33593	159	24	models	model	NOUN
bracis-33593	159	25	learned	learn	VERB
bracis-33593	159	26	trough	trough	NOUN
bracis-33593	159	27	a	a	DET
bracis-33593	159	28	episode	episode	NOUN
bracis-33593	159	29	,	,	PUNCT
bracis-33593	159	30	on	on	ADP
bracis-33593	159	31	the	the	DET
bracis-33593	159	32	left	left	NOUN
bracis-33593	159	33	is	be	AUX
bracis-33593	159	34	the	the	DET
bracis-33593	159	35	tree	tree	NOUN
bracis-33593	159	36	generated	generate	VERB
bracis-33593	159	37	in	in	ADP
bracis-33593	159	38	the	the	DET
bracis-33593	159	39	beginning	beginning	NOUN
bracis-33593	159	40	and	and	CCONJ
bracis-33593	159	41	in	in	ADP
bracis-33593	159	42	the	the	DET
bracis-33593	159	43	right	right	NOUN
bracis-33593	159	44	at	at	ADP
bracis-33593	159	45	the	the	DET
bracis-33593	159	46	end	end	NOUN
bracis-33593	159	47	of	of	ADP
bracis-33593	159	48	the	the	DET
bracis-33593	159	49	episode	episode	NOUN
bracis-33593	159	50	.	.	PUNCT
bracis-33593	160	1	we	we	PRON
bracis-33593	160	2	can	can	AUX
bracis-33593	160	3	see	see	VERB
bracis-33593	160	4	how	how	SCONJ
bracis-33593	160	5	in	in	ADP
bracis-33593	160	6	the	the	DET
bracis-33593	160	7	end	end	NOUN
bracis-33593	160	8	of	of	ADP
bracis-33593	160	9	the	the	DET
bracis-33593	160	10	episode	episode	NOUN
bracis-33593	160	11	,	,	PUNCT
bracis-33593	160	12	the	the	DET
bracis-33593	160	13	quality	quality	NOUN
bracis-33593	160	14	of	of	ADP
bracis-33593	160	15	the	the	DET
bracis-33593	160	16	models	model	NOUN
bracis-33593	160	17	were	be	AUX
bracis-33593	160	18	improved	improve	VERB
bracis-33593	160	19	resulting	result	VERB
bracis-33593	160	20	in	in	ADP
bracis-33593	160	21	a	a	DET
bracis-33593	160	22	more	more	ADV
bracis-33593	160	23	consistent	consistent	ADJ
bracis-33593	160	24	plan	plan	NOUN
bracis-33593	160	25	.	.	PUNCT
bracis-33593	161	1	6	6	NUM
bracis-33593	161	2	results	result	NOUN
bracis-33593	161	3	to	to	PART
bracis-33593	161	4	compare	compare	VERB
bracis-33593	161	5	the	the	DET
bracis-33593	161	6	advantages	advantage	NOUN
bracis-33593	161	7	of	of	ADP
bracis-33593	161	8	using	use	VERB
bracis-33593	161	9	a	a	DET
bracis-33593	161	10	segmented	segment	VERB
bracis-33593	161	11	representation	representation	NOUN
bracis-33593	161	12	of	of	ADP
bracis-33593	161	13	the	the	DET
bracis-33593	161	14	world	world	NOUN
bracis-33593	161	15	against	against	ADP
bracis-33593	161	16	the	the	DET
bracis-33593	161	17	standard	standard	ADJ
bracis-33593	161	18	approach	approach	NOUN
bracis-33593	161	19	of	of	ADP
bracis-33593	161	20	a	a	DET
bracis-33593	161	21	single	single	ADJ
bracis-33593	161	22	unified	unified	ADJ
bracis-33593	161	23	model	model	NOUN
bracis-33593	161	24	,	,	PUNCT
bracis-33593	161	25	we	we	PRON
bracis-33593	161	26	designed	design	VERB
bracis-33593	161	27	a	a	DET
bracis-33593	161	28	series	series	NOUN
bracis-33593	161	29	of	of	ADP
bracis-33593	161	30	experiments	experiment	NOUN
bracis-33593	161	31	aimed	aim	VERB
bracis-33593	161	32	at	at	ADP
bracis-33593	161	33	highlighting	highlight	VERB
bracis-33593	161	34	the	the	DET
bracis-33593	161	35	differences	difference	NOUN
bracis-33593	161	36	in	in	ADP
bracis-33593	161	37	performance	performance	NOUN
bracis-33593	161	38	.	.	PUNCT
bracis-33593	162	1	first	first	ADV
bracis-33593	162	2	we	we	PRON
bracis-33593	162	3	design	design	VERB
bracis-33593	162	4	the	the	DET
bracis-33593	162	5	experiment	experiment	NOUN
bracis-33593	162	6	in	in	ADP
bracis-33593	162	7	which	which	PRON
bracis-33593	162	8	we	we	PRON
bracis-33593	162	9	conducted	conduct	VERB
bracis-33593	162	10	ten	ten	NUM
bracis-33593	162	11	runs	run	NOUN
bracis-33593	162	12	of	of	ADP
bracis-33593	162	13	two	two	NUM
bracis-33593	162	14	episodes	episode	NOUN
bracis-33593	162	15	with	with	ADP
bracis-33593	162	16	at	at	ADP
bracis-33593	162	17	most	most	ADV
bracis-33593	162	18	three	three	NUM
bracis-33593	162	19	hundred	hundred	NUM
bracis-33593	162	20	steps	step	NOUN
bracis-33593	162	21	each	each	PRON
bracis-33593	162	22	,	,	PUNCT
bracis-33593	162	23	where	where	SCONJ
bracis-33593	162	24	both	both	DET
bracis-33593	162	25	models	model	NOUN
bracis-33593	162	26	interacted	interact	VERB
bracis-33593	162	27	with	with	ADP
bracis-33593	162	28	the	the	DET
bracis-33593	162	29	same	same	ADJ
bracis-33593	162	30	environment	environment	NOUN
bracis-33593	162	31	.	.	PUNCT
bracis-33593	163	1	by	by	ADP
bracis-33593	163	2	comparing	compare	VERB
bracis-33593	163	3	their	their	PRON
bracis-33593	163	4	performances	performance	NOUN
bracis-33593	163	5	in	in	ADP
bracis-33593	163	6	this	this	DET
bracis-33593	163	7	controlled	control	VERB
bracis-33593	163	8	setting	setting	NOUN
bracis-33593	163	9	,	,	PUNCT
bracis-33593	163	10	we	we	PRON
bracis-33593	163	11	expect	expect	VERB
bracis-33593	163	12	to	to	PART
bracis-33593	163	13	have	have	VERB
bracis-33593	163	14	both	both	DET
bracis-33593	163	15	solution	solution	NOUN
bracis-33593	163	16	being	be	AUX
bracis-33593	163	17	very	very	ADV
bracis-33593	163	18	similar	similar	ADJ
bracis-33593	163	19	in	in	ADP
bracis-33593	163	20	the	the	DET
bracis-33593	163	21	first	first	ADJ
bracis-33593	163	22	episode	episode	NOUN
bracis-33593	163	23	.	.	PUNCT
bracis-33593	164	1	we	we	PRON
bracis-33593	164	2	limited	limit	VERB
bracis-33593	164	3	the	the	DET
bracis-33593	164	4	horizon	horizon	NOUN
bracis-33593	164	5	to	to	ADP
bracis-33593	164	6	three	three	NUM
bracis-33593	164	7	hundred	hundred	NUM
bracis-33593	164	8	steps	step	NOUN
bracis-33593	164	9	.	.	PUNCT
bracis-33593	165	1	so	so	ADV
bracis-33593	165	2	with	with	ADP
bracis-33593	165	3	this	this	DET
bracis-33593	165	4	experiment	experiment	NOUN
bracis-33593	165	5	setup	setup	NOUN
bracis-33593	165	6	we	we	PRON
bracis-33593	165	7	develop	develop	VERB
bracis-33593	165	8	two	two	NUM
bracis-33593	165	9	scenarios	scenario	NOUN
bracis-33593	165	10	,	,	PUNCT
bracis-33593	165	11	the	the	DET
bracis-33593	165	12	first	first	ADJ
bracis-33593	165	13	one	one	NUM
bracis-33593	165	14	just	just	ADV
bracis-33593	165	15	as	as	SCONJ
bracis-33593	165	16	we	we	PRON
bracis-33593	165	17	described	describe	VERB
bracis-33593	165	18	but	but	CCONJ
bracis-33593	165	19	the	the	DET
bracis-33593	165	20	second	second	ADJ
bracis-33593	165	21	one	one	NOUN
bracis-33593	165	22	we	we	PRON
bracis-33593	165	23	change	change	VERB
bracis-33593	165	24	the	the	DET
bracis-33593	165	25	environment	environment	NOUN
bracis-33593	165	26	between	between	ADP
bracis-33593	165	27	first	first	ADJ
bracis-33593	165	28	and	and	CCONJ
bracis-33593	165	29	second	second	ADJ
bracis-33593	165	30	episode	episode	NOUN
bracis-33593	165	31	.	.	PUNCT
bracis-33593	166	1	as	as	SCONJ
bracis-33593	166	2	the	the	DET
bracis-33593	166	3	agent	agent	NOUN
bracis-33593	166	4	maintain	maintain	VERB
bracis-33593	166	5	the	the	DET
bracis-33593	166	6	knowledge	knowledge	NOUN
bracis-33593	166	7	of	of	ADP
bracis-33593	166	8	the	the	DET
bracis-33593	166	9	previous	previous	ADJ
bracis-33593	166	10	episode	episode	NOUN
bracis-33593	166	11	,	,	PUNCT
bracis-33593	166	12	we	we	PRON
bracis-33593	166	13	expected	expect	VERB
bracis-33593	166	14	to	to	PART
bracis-33593	166	15	see	see	VERB
bracis-33593	166	16	a	a	DET
bracis-33593	166	17	better	well	ADJ
bracis-33593	166	18	adaption	adaption	NOUN
bracis-33593	166	19	by	by	ADP
bracis-33593	166	20	the	the	DET
bracis-33593	166	21	k	k	ADJ
bracis-33593	166	22	-	-	PUNCT
bracis-33593	166	23	model	model	NOUN
bracis-33593	166	24	solution	solution	NOUN
bracis-33593	166	25	in	in	ADP
bracis-33593	166	26	a	a	DET
bracis-33593	166	27	different	different	ADJ
bracis-33593	166	28	environment	environment	NOUN
bracis-33593	166	29	that	that	PRON
bracis-33593	166	30	brings	bring	VERB
bracis-33593	166	31	familiar	familiar	ADJ
bracis-33593	166	32	dynamics	dynamic	NOUN
bracis-33593	166	33	.	.	PUNCT
bracis-33593	167	1	so	so	ADV
bracis-33593	167	2	,	,	PUNCT
bracis-33593	167	3	for	for	ADP
bracis-33593	167	4	the	the	DET
bracis-33593	167	5	first	first	ADJ
bracis-33593	167	6	experiment	experiment	NOUN
bracis-33593	167	7	,	,	PUNCT
bracis-33593	167	8	we	we	PRON
bracis-33593	167	9	run	run	VERB
bracis-33593	167	10	the	the	DET
bracis-33593	167	11	process	process	NOUN
bracis-33593	167	12	into	into	ADP
bracis-33593	167	13	two	two	NUM
bracis-33593	167	14	episodes	episode	NOUN
bracis-33593	167	15	.	.	PUNCT
bracis-33593	168	1	in	in	ADP
bracis-33593	168	2	this	this	DET
bracis-33593	168	3	setup	setup	NOUN
bracis-33593	168	4	,	,	PUNCT
bracis-33593	168	5	each	each	DET
bracis-33593	168	6	solution	solution	NOUN
bracis-33593	168	7	retained	retain	VERB
bracis-33593	168	8	knowledge	knowledge	NOUN
bracis-33593	168	9	from	from	ADP
bracis-33593	168	10	previous	previous	ADJ
bracis-33593	168	11	episodes	episode	NOUN
bracis-33593	168	12	,	,	PUNCT
bracis-33593	168	13	allowing	allow	VERB
bracis-33593	168	14	us	we	PRON
bracis-33593	168	15	to	to	PART
bracis-33593	168	16	observe	observe	VERB
bracis-33593	168	17	how	how	SCONJ
bracis-33593	168	18	their	their	PRON
bracis-33593	168	19	performance	performance	NOUN
bracis-33593	168	20	evolved	evolve	VERB
bracis-33593	168	21	over	over	ADP
bracis-33593	168	22	time	time	NOUN
bracis-33593	168	23	.	.	PUNCT
bracis-33593	169	1	we	we	PRON
bracis-33593	169	2	expected	expect	VERB
bracis-33593	169	3	both	both	DET
bracis-33593	169	4	models	model	NOUN
bracis-33593	169	5	to	to	PART
bracis-33593	169	6	improve	improve	VERB
bracis-33593	169	7	across	across	ADP
bracis-33593	169	8	episodes	episode	NOUN
bracis-33593	169	9	,	,	PUNCT
bracis-33593	169	10	leveraging	leverage	VERB
bracis-33593	169	11	their	their	PRON
bracis-33593	169	12	accumulated	accumulate	VERB
bracis-33593	169	13	experience	experience	NOUN
bracis-33593	169	14	to	to	PART
bracis-33593	169	15	make	make	VERB
bracis-33593	169	16	increasingly	increasingly	ADV
bracis-33593	169	17	informed	inform	VERB
bracis-33593	169	18	decisions	decision	NOUN
bracis-33593	169	19	.	.	PUNCT
bracis-33593	170	1	fig	fig	NOUN
bracis-33593	170	2	.	.	PUNCT
bracis-33593	171	1	5	5	X
bracis-33593	171	2	.	.	X
bracis-33593	171	3	the	the	DET
bracis-33593	171	4	different	different	ADJ
bracis-33593	171	5	environment	environment	NOUN
bracis-33593	171	6	presented	present	VERB
bracis-33593	171	7	in	in	ADP
bracis-33593	171	8	scenario	scenario	NOUN
bracis-33593	171	9	2	2	NUM
bracis-33593	171	10	where	where	SCONJ
bracis-33593	171	11	the	the	DET
bracis-33593	171	12	environment	environment	NOUN
bracis-33593	171	13	changes	change	VERB
bracis-33593	171	14	between	between	ADP
bracis-33593	171	15	episodes	episode	NOUN
bracis-33593	171	16	.	.	PUNCT
bracis-33593	172	1	full	full	ADJ
bracis-33593	172	2	size	size	NOUN
bracis-33593	172	3	image	image	NOUN
bracis-33593	172	4	in	in	ADP
bracis-33593	172	5	the	the	DET
bracis-33593	172	6	second	second	ADJ
bracis-33593	172	7	experiment	experiment	NOUN
bracis-33593	172	8	,	,	PUNCT
bracis-33593	172	9	we	we	PRON
bracis-33593	172	10	introduced	introduce	VERB
bracis-33593	172	11	a	a	DET
bracis-33593	172	12	novel	novel	ADJ
bracis-33593	172	13	twist	twist	NOUN
bracis-33593	172	14	by	by	ADP
bracis-33593	172	15	presenting	present	VERB
bracis-33593	172	16	the	the	DET
bracis-33593	172	17	agent	agent	NOUN
bracis-33593	172	18	with	with	ADP
bracis-33593	172	19	a	a	DET
bracis-33593	172	20	different	different	ADJ
bracis-33593	172	21	environment	environment	NOUN
bracis-33593	172	22	in	in	ADP
bracis-33593	172	23	the	the	DET
bracis-33593	172	24	second	second	ADJ
bracis-33593	172	25	episode	episode	NOUN
bracis-33593	172	26	.	.	PUNCT
bracis-33593	173	1	while	while	SCONJ
bracis-33593	173	2	the	the	DET
bracis-33593	173	3	dynamics	dynamic	NOUN
bracis-33593	173	4	of	of	ADP
bracis-33593	173	5	this	this	DET
bracis-33593	173	6	environment	environment	NOUN
bracis-33593	173	7	remained	remain	VERB
bracis-33593	173	8	very	very	ADV
bracis-33593	173	9	similar	similar	ADJ
bracis-33593	173	10	to	to	ADP
bracis-33593	173	11	the	the	DET
bracis-33593	173	12	one	one	NOUN
bracis-33593	173	13	previously	previously	ADV
bracis-33593	173	14	encountered	encounter	VERB
bracis-33593	173	15	,	,	PUNCT
bracis-33593	173	16	the	the	DET
bracis-33593	173	17	regions	region	NOUN
bracis-33593	173	18	within	within	ADP
bracis-33593	173	19	it	it	PRON
bracis-33593	173	20	were	be	AUX
bracis-33593	173	21	arranged	arrange	VERB
bracis-33593	173	22	in	in	ADP
bracis-33593	173	23	different	different	ADJ
bracis-33593	173	24	shapes	shape	NOUN
bracis-33593	173	25	and	and	CCONJ
bracis-33593	173	26	positions	position	NOUN
bracis-33593	173	27	,	,	PUNCT
bracis-33593	173	28	as	as	SCONJ
bracis-33593	173	29	depicted	depict	VERB
bracis-33593	173	30	in	in	ADP
bracis-33593	173	31	fig	fig	NOUN
bracis-33593	173	32	.	.	PUNCT
bracis-33593	173	33	 	 	SPACE
bracis-33593	174	1	5	5	NUM
bracis-33593	174	2	.	.	PUNCT
bracis-33593	175	1	despite	despite	SCONJ
bracis-33593	175	2	this	this	DET
bracis-33593	175	3	change	change	NOUN
bracis-33593	175	4	,	,	PUNCT
bracis-33593	175	5	both	both	DET
bracis-33593	175	6	models	model	NOUN
bracis-33593	175	7	retained	retain	VERB
bracis-33593	175	8	persistent	persistent	ADJ
bracis-33593	175	9	knowledge	knowledge	NOUN
bracis-33593	175	10	from	from	ADP
bracis-33593	175	11	the	the	DET
bracis-33593	175	12	previous	previous	ADJ
bracis-33593	175	13	episodes	episode	NOUN
bracis-33593	175	14	.	.	PUNCT
bracis-33593	176	1	however	however	ADV
bracis-33593	176	2	,	,	PUNCT
bracis-33593	176	3	we	we	PRON
bracis-33593	176	4	anticipated	anticipate	VERB
bracis-33593	176	5	that	that	SCONJ
bracis-33593	176	6	the	the	DET
bracis-33593	176	7	segmented	segment	VERB
bracis-33593	176	8	model	model	NOUN
bracis-33593	176	9	would	would	AUX
bracis-33593	176	10	exhibit	exhibit	VERB
bracis-33593	176	11	superior	superior	ADJ
bracis-33593	176	12	adaptability	adaptability	NOUN
bracis-33593	176	13	in	in	ADP
bracis-33593	176	14	this	this	DET
bracis-33593	176	15	scenario	scenario	NOUN
bracis-33593	176	16	.	.	PUNCT
bracis-33593	177	1	this	this	DET
bracis-33593	177	2	expectation	expectation	NOUN
bracis-33593	177	3	stemmed	stem	VERB
bracis-33593	177	4	from	from	ADP
bracis-33593	177	5	the	the	DET
bracis-33593	177	6	segmented	segment	VERB
bracis-33593	177	7	model	model	NOUN
bracis-33593	177	8	’s	’s	PART
bracis-33593	177	9	ability	ability	NOUN
bracis-33593	177	10	to	to	PART
bracis-33593	177	11	identify	identify	VERB
bracis-33593	177	12	previously	previously	ADV
bracis-33593	177	13	learned	learn	VERB
bracis-33593	177	14	models	model	NOUN
bracis-33593	177	15	and	and	CCONJ
bracis-33593	177	16	adjust	adjust	VERB
bracis-33593	177	17	its	its	PRON
bracis-33593	177	18	routing	routing	NOUN
bracis-33593	177	19	network	network	NOUN
bracis-33593	177	20	accordingly	accordingly	ADV
bracis-33593	177	21	to	to	PART
bracis-33593	177	22	accommodate	accommodate	VERB
bracis-33593	177	23	the	the	DET
bracis-33593	177	24	new	new	ADJ
bracis-33593	177	25	configuration	configuration	NOUN
bracis-33593	177	26	of	of	ADP
bracis-33593	177	27	regions	region	NOUN
bracis-33593	177	28	.	.	PUNCT
bracis-33593	178	1	fig	fig	NOUN
bracis-33593	178	2	.	.	PUNCT
bracis-33593	179	1	6	6	NUM
bracis-33593	179	2	.	.	X
bracis-33593	179	3	box	box	NOUN
bracis-33593	179	4	plot	plot	NOUN
bracis-33593	179	5	showing	show	VERB
bracis-33593	179	6	distribution	distribution	NOUN
bracis-33593	179	7	of	of	ADP
bracis-33593	179	8	steps	step	NOUN
bracis-33593	179	9	that	that	PRON
bracis-33593	179	10	each	each	DET
bracis-33593	179	11	model	model	NOUN
bracis-33593	179	12	spent	spend	VERB
bracis-33593	179	13	to	to	PART
bracis-33593	179	14	get	get	VERB
bracis-33593	179	15	to	to	ADP
bracis-33593	179	16	the	the	DET
bracis-33593	179	17	goal	goal	NOUN
bracis-33593	179	18	across	across	ADP
bracis-33593	179	19	first	first	ADJ
bracis-33593	179	20	episode	episode	NOUN
bracis-33593	179	21	of	of	ADP
bracis-33593	179	22	the	the	DET
bracis-33593	179	23	first	first	ADJ
bracis-33593	179	24	scenario	scenario	NOUN
bracis-33593	179	25	with	with	ADP
bracis-33593	179	26	blue	blue	ADJ
bracis-33593	179	27	being	be	AUX
bracis-33593	179	28	the	the	DET
bracis-33593	179	29	standard	standard	ADJ
bracis-33593	179	30	single	single	ADJ
bracis-33593	179	31	model	model	NOUN
bracis-33593	179	32	approach	approach	NOUN
bracis-33593	179	33	and	and	CCONJ
bracis-33593	179	34	orange	orange	NOUN
bracis-33593	179	35	being	be	AUX
bracis-33593	179	36	the	the	DET
bracis-33593	179	37	k	k	NOUN
bracis-33593	179	38	-	-	PUNCT
bracis-33593	179	39	model	model	NOUN
bracis-33593	179	40	segmented	segment	VERB
bracis-33593	179	41	approach	approach	NOUN
bracis-33593	179	42	.	.	PUNCT
bracis-33593	180	1	(	(	PUNCT
bracis-33593	180	2	color	color	NOUN
bracis-33593	180	3	figure	figure	NOUN
bracis-33593	180	4	online	online	ADV
bracis-33593	180	5	)	)	PUNCT
bracis-33593	180	6	full	full	ADJ
bracis-33593	180	7	size	size	NOUN
bracis-33593	180	8	image	image	NOUN
bracis-33593	180	9	the	the	DET
bracis-33593	180	10	analysis	analysis	NOUN
bracis-33593	180	11	presented	present	VERB
bracis-33593	180	12	in	in	ADP
bracis-33593	180	13	fig	fig	NOUN
bracis-33593	180	14	.	.	PUNCT
bracis-33593	180	15	 	 	SPACE
bracis-33593	180	16	6	6	NUM
bracis-33593	180	17	reveals	reveal	VERB
bracis-33593	180	18	interesting	interesting	ADJ
bracis-33593	180	19	insights	insight	NOUN
bracis-33593	180	20	into	into	ADP
bracis-33593	180	21	the	the	DET
bracis-33593	180	22	performance	performance	NOUN
bracis-33593	180	23	of	of	ADP
bracis-33593	180	24	both	both	DET
bracis-33593	180	25	approaches	approach	NOUN
bracis-33593	180	26	across	across	ADP
bracis-33593	180	27	episodes	episode	NOUN
bracis-33593	180	28	.	.	PUNCT
bracis-33593	181	1	initially	initially	ADV
bracis-33593	181	2	,	,	PUNCT
bracis-33593	181	3	in	in	ADP
bracis-33593	181	4	the	the	DET
bracis-33593	181	5	first	first	ADJ
bracis-33593	181	6	episode	episode	NOUN
bracis-33593	181	7	,	,	PUNCT
bracis-33593	181	8	both	both	CCONJ
bracis-33593	181	9	the	the	DET
bracis-33593	181	10	single	single	ADJ
bracis-33593	181	11	-	-	PUNCT
bracis-33593	181	12	model	model	NOUN
bracis-33593	181	13	and	and	CCONJ
bracis-33593	181	14	the	the	DET
bracis-33593	181	15	k	k	NOUN
bracis-33593	181	16	-	-	PUNCT
bracis-33593	181	17	models	model	NOUN
bracis-33593	181	18	exhibit	exhibit	VERB
bracis-33593	181	19	comparable	comparable	ADJ
bracis-33593	181	20	performances	performance	NOUN
bracis-33593	181	21	,	,	PUNCT
bracis-33593	181	22	as	as	SCONJ
bracis-33593	181	23	evidenced	evidence	VERB
bracis-33593	181	24	by	by	ADP
bracis-33593	181	25	the	the	DET
bracis-33593	181	26	similar	similar	ADJ
bracis-33593	181	27	duration	duration	NOUN
bracis-33593	181	28	of	of	ADP
bracis-33593	181	29	the	the	DET
bracis-33593	181	30	episodes	episode	NOUN
bracis-33593	181	31	with	with	ADP
bracis-33593	181	32	the	the	DET
bracis-33593	181	33	k	k	NOUN
bracis-33593	181	34	-	-	NOUN
bracis-33593	181	35	model	model	NOUN
bracis-33593	181	36	being	be	AUX
bracis-33593	181	37	slightly	slightly	ADV
bracis-33593	181	38	worse	bad	ADJ
bracis-33593	181	39	.	.	PUNCT
bracis-33593	182	1	this	this	PRON
bracis-33593	182	2	suggests	suggest	VERB
bracis-33593	182	3	that	that	SCONJ
bracis-33593	182	4	both	both	DET
bracis-33593	182	5	models	model	NOUN
bracis-33593	182	6	are	be	AUX
bracis-33593	182	7	capable	capable	ADJ
bracis-33593	182	8	of	of	ADP
bracis-33593	182	9	effectively	effectively	ADV
bracis-33593	182	10	learning	learn	VERB
bracis-33593	182	11	to	to	PART
bracis-33593	182	12	navigate	navigate	VERB
bracis-33593	182	13	through	through	ADP
bracis-33593	182	14	the	the	DET
bracis-33593	182	15	environment	environment	NOUN
bracis-33593	182	16	in	in	ADP
bracis-33593	182	17	a	a	DET
bracis-33593	182	18	similar	similar	ADJ
bracis-33593	182	19	fashion	fashion	NOUN
bracis-33593	182	20	.	.	PUNCT
bracis-33593	183	1	however	however	ADV
bracis-33593	183	2	,	,	PUNCT
bracis-33593	183	3	in	in	ADP
bracis-33593	183	4	the	the	DET
bracis-33593	183	5	subsequent	subsequent	ADJ
bracis-33593	183	6	episode	episode	NOUN
bracis-33593	183	7	shown	show	VERB
bracis-33593	183	8	in	in	ADP
bracis-33593	183	9	fig	fig	NOUN
bracis-33593	183	10	.	.	PUNCT
bracis-33593	183	11	 	 	SPACE
bracis-33593	183	12	6	6	NUM
bracis-33593	183	13	a	a	DET
bracis-33593	183	14	notable	notable	ADJ
bracis-33593	183	15	disparity	disparity	NOUN
bracis-33593	183	16	emerges	emerge	VERB
bracis-33593	183	17	between	between	ADP
bracis-33593	183	18	the	the	DET
bracis-33593	183	19	two	two	NUM
bracis-33593	183	20	approaches	approach	NOUN
bracis-33593	183	21	.	.	PUNCT
bracis-33593	184	1	the	the	DET
bracis-33593	184	2	k	k	NOUN
bracis-33593	184	3	-	-	PUNCT
bracis-33593	184	4	models	model	NOUN
bracis-33593	184	5	demonstrate	demonstrate	VERB
bracis-33593	184	6	a	a	DET
bracis-33593	184	7	significant	significant	ADJ
bracis-33593	184	8	reduction	reduction	NOUN
bracis-33593	184	9	in	in	ADP
bracis-33593	184	10	episode	episode	NOUN
bracis-33593	184	11	length	length	NOUN
bracis-33593	184	12	in	in	ADP
bracis-33593	184	13	most	most	ADJ
bracis-33593	184	14	of	of	ADP
bracis-33593	184	15	the	the	DET
bracis-33593	184	16	runs	run	NOUN
bracis-33593	184	17	,	,	PUNCT
bracis-33593	184	18	compared	compare	VERB
bracis-33593	184	19	to	to	ADP
bracis-33593	184	20	the	the	DET
bracis-33593	184	21	single	single	ADJ
bracis-33593	184	22	model	model	NOUN
bracis-33593	184	23	.	.	PUNCT
bracis-33593	185	1	this	this	DET
bracis-33593	185	2	observation	observation	NOUN
bracis-33593	185	3	suggests	suggest	VERB
bracis-33593	185	4	that	that	SCONJ
bracis-33593	185	5	the	the	DET
bracis-33593	185	6	k	k	NOUN
bracis-33593	185	7	-	-	PUNCT
bracis-33593	185	8	models	model	NOUN
bracis-33593	185	9	have	have	VERB
bracis-33593	185	10	a	a	DET
bracis-33593	185	11	better	well	ADJ
bracis-33593	185	12	reset	reset	ADJ
bracis-33593	185	13	power	power	NOUN
bracis-33593	185	14	to	to	PART
bracis-33593	185	15	retrieve	retrieve	VERB
bracis-33593	185	16	the	the	DET
bracis-33593	185	17	knowledge	knowledge	NOUN
bracis-33593	185	18	from	from	ADP
bracis-33593	185	19	the	the	DET
bracis-33593	185	20	beginning	beginning	NOUN
bracis-33593	185	21	of	of	ADP
bracis-33593	185	22	the	the	DET
bracis-33593	185	23	previous	previous	ADJ
bracis-33593	185	24	episode	episode	NOUN
bracis-33593	185	25	,	,	PUNCT
bracis-33593	185	26	while	while	SCONJ
bracis-33593	185	27	the	the	DET
bracis-33593	185	28	changes	change	NOUN
bracis-33593	185	29	presented	present	VERB
bracis-33593	185	30	through	through	ADP
bracis-33593	185	31	the	the	DET
bracis-33593	185	32	episode	episode	NOUN
bracis-33593	185	33	had	have	VERB
bracis-33593	185	34	more	more	ADJ
bracis-33593	185	35	impact	impact	NOUN
bracis-33593	185	36	in	in	ADP
bracis-33593	185	37	the	the	DET
bracis-33593	185	38	single	single	ADJ
bracis-33593	185	39	model	model	NOUN
bracis-33593	185	40	approach	approach	NOUN
bracis-33593	185	41	,	,	PUNCT
bracis-33593	185	42	resulting	result	VERB
bracis-33593	185	43	in	in	ADP
bracis-33593	185	44	a	a	DET
bracis-33593	185	45	better	well	ADJ
bracis-33593	185	46	performance	performance	NOUN
bracis-33593	185	47	compared	compare	VERB
bracis-33593	185	48	to	to	ADP
bracis-33593	185	49	that	that	PRON
bracis-33593	185	50	observed	observe	VERB
bracis-33593	185	51	in	in	ADP
bracis-33593	185	52	the	the	DET
bracis-33593	185	53	first	first	ADJ
bracis-33593	185	54	episode	episode	NOUN
bracis-33593	185	55	but	but	CCONJ
bracis-33593	185	56	not	not	PART
bracis-33593	185	57	as	as	ADV
bracis-33593	185	58	good	good	ADJ
bracis-33593	185	59	as	as	ADP
bracis-33593	185	60	the	the	DET
bracis-33593	185	61	improvement	improvement	NOUN
bracis-33593	185	62	observed	observe	VERB
bracis-33593	185	63	by	by	ADP
bracis-33593	185	64	the	the	DET
bracis-33593	185	65	k	k	NOUN
bracis-33593	185	66	-	-	NOUN
bracis-33593	185	67	models	model	NOUN
bracis-33593	185	68	.	.	PUNCT
bracis-33593	186	1	the	the	DET
bracis-33593	186	2	improvement	improvement	NOUN
bracis-33593	186	3	became	become	VERB
bracis-33593	186	4	even	even	ADV
bracis-33593	186	5	more	more	ADV
bracis-33593	186	6	evident	evident	ADJ
bracis-33593	186	7	when	when	SCONJ
bracis-33593	186	8	analysing	analyse	VERB
bracis-33593	186	9	the	the	DET
bracis-33593	186	10	performances	performance	NOUN
bracis-33593	186	11	in	in	ADP
bracis-33593	186	12	the	the	DET
bracis-33593	186	13	second	second	ADJ
bracis-33593	186	14	episode	episode	NOUN
bracis-33593	186	15	for	for	ADP
bracis-33593	186	16	the	the	DET
bracis-33593	186	17	scenario	scenario	NOUN
bracis-33593	186	18	when	when	SCONJ
bracis-33593	186	19	the	the	DET
bracis-33593	186	20	environment	environment	NOUN
bracis-33593	186	21	changes	change	VERB
bracis-33593	186	22	.	.	PUNCT
bracis-33593	187	1	the	the	DET
bracis-33593	187	2	fig	fig	NOUN
bracis-33593	187	3	.	.	PUNCT
bracis-33593	187	4	 	 	SPACE
bracis-33593	187	5	7	7	NUM
bracis-33593	187	6	show	show	VERB
bracis-33593	187	7	us	we	PRON
bracis-33593	187	8	the	the	DET
bracis-33593	187	9	difference	difference	NOUN
bracis-33593	187	10	between	between	ADP
bracis-33593	187	11	the	the	DET
bracis-33593	187	12	two	two	NUM
bracis-33593	187	13	approaches	approach	NOUN
bracis-33593	187	14	on	on	ADP
bracis-33593	187	15	this	this	DET
bracis-33593	187	16	scenario	scenario	NOUN
bracis-33593	187	17	.	.	PUNCT
bracis-33593	188	1	the	the	DET
bracis-33593	188	2	k	k	NOUN
bracis-33593	188	3	-	-	PUNCT
bracis-33593	188	4	models	model	NOUN
bracis-33593	188	5	demonstrate	demonstrate	VERB
bracis-33593	188	6	a	a	DET
bracis-33593	188	7	significant	significant	ADJ
bracis-33593	188	8	reduction	reduction	NOUN
bracis-33593	188	9	in	in	ADP
bracis-33593	188	10	episode	episode	NOUN
bracis-33593	188	11	length	length	NOUN
bracis-33593	188	12	in	in	ADP
bracis-33593	188	13	most	most	ADJ
bracis-33593	188	14	of	of	ADP
bracis-33593	188	15	the	the	DET
bracis-33593	188	16	runs	run	NOUN
bracis-33593	188	17	,	,	PUNCT
bracis-33593	188	18	compared	compare	VERB
bracis-33593	188	19	to	to	ADP
bracis-33593	188	20	the	the	DET
bracis-33593	188	21	single	single	ADJ
bracis-33593	188	22	model	model	NOUN
bracis-33593	188	23	as	as	SCONJ
bracis-33593	188	24	it	it	PRON
bracis-33593	188	25	did	do	VERB
bracis-33593	188	26	in	in	ADP
bracis-33593	188	27	the	the	DET
bracis-33593	188	28	first	first	ADJ
bracis-33593	188	29	scenario	scenario	NOUN
bracis-33593	188	30	,	,	PUNCT
bracis-33593	188	31	but	but	CCONJ
bracis-33593	188	32	this	this	DET
bracis-33593	188	33	time	time	NOUN
bracis-33593	188	34	we	we	PRON
bracis-33593	188	35	ca	can	AUX
bracis-33593	188	36	n’t	not	PART
bracis-33593	188	37	see	see	VERB
bracis-33593	188	38	the	the	DET
bracis-33593	188	39	same	same	ADJ
bracis-33593	188	40	improvement	improvement	NOUN
bracis-33593	188	41	in	in	ADP
bracis-33593	188	42	the	the	DET
bracis-33593	188	43	single	single	ADJ
bracis-33593	188	44	model	model	NOUN
bracis-33593	188	45	approach	approach	NOUN
bracis-33593	188	46	.	.	PUNCT
bracis-33593	189	1	this	this	DET
bracis-33593	189	2	observation	observation	NOUN
bracis-33593	189	3	suggests	suggest	VERB
bracis-33593	189	4	that	that	SCONJ
bracis-33593	189	5	the	the	DET
bracis-33593	189	6	single	single	ADJ
bracis-33593	189	7	model	model	NOUN
bracis-33593	189	8	struggles	struggle	VERB
bracis-33593	189	9	to	to	PART
bracis-33593	189	10	transfer	transfer	VERB
bracis-33593	189	11	knowledge	knowledge	NOUN
bracis-33593	189	12	from	from	ADP
bracis-33593	189	13	the	the	DET
bracis-33593	189	14	previous	previous	ADJ
bracis-33593	189	15	episode	episode	NOUN
bracis-33593	189	16	to	to	ADP
bracis-33593	189	17	the	the	DET
bracis-33593	189	18	next	next	ADJ
bracis-33593	189	19	when	when	SCONJ
bracis-33593	189	20	the	the	DET
bracis-33593	189	21	environment	environment	NOUN
bracis-33593	189	22	changes	change	VERB
bracis-33593	189	23	,	,	PUNCT
bracis-33593	189	24	resulting	result	VERB
bracis-33593	189	25	in	in	ADP
bracis-33593	189	26	performance	performance	NOUN
bracis-33593	189	27	similar	similar	ADJ
bracis-33593	189	28	to	to	ADP
bracis-33593	189	29	that	that	PRON
bracis-33593	189	30	observed	observe	VERB
bracis-33593	189	31	in	in	ADP
bracis-33593	189	32	the	the	DET
bracis-33593	189	33	first	first	ADJ
bracis-33593	189	34	episode	episode	NOUN
bracis-33593	189	35	,	,	PUNCT
bracis-33593	189	36	where	where	SCONJ
bracis-33593	189	37	it	it	PRON
bracis-33593	189	38	had	have	VERB
bracis-33593	189	39	to	to	PART
bracis-33593	189	40	learn	learn	VERB
bracis-33593	189	41	the	the	DET
bracis-33593	189	42	environment	environment	NOUN
bracis-33593	189	43	from	from	ADP
bracis-33593	189	44	scratch	scratch	NOUN
bracis-33593	189	45	.	.	PUNCT
bracis-33593	190	1	in	in	ADP
bracis-33593	190	2	contrast	contrast	NOUN
bracis-33593	190	3	,	,	PUNCT
bracis-33593	190	4	the	the	DET
bracis-33593	190	5	k	k	NOUN
bracis-33593	190	6	-	-	PUNCT
bracis-33593	190	7	models	model	NOUN
bracis-33593	190	8	leverage	leverage	VERB
bracis-33593	190	9	the	the	DET
bracis-33593	190	10	dynamics	dynamic	NOUN
bracis-33593	190	11	learned	learn	VERB
bracis-33593	190	12	in	in	ADP
bracis-33593	190	13	the	the	DET
bracis-33593	190	14	first	first	ADJ
bracis-33593	190	15	episode	episode	NOUN
bracis-33593	190	16	to	to	PART
bracis-33593	190	17	adapt	adapt	VERB
bracis-33593	190	18	and	and	CCONJ
bracis-33593	190	19	improve	improve	VERB
bracis-33593	190	20	their	their	PRON
bracis-33593	190	21	performance	performance	NOUN
bracis-33593	190	22	in	in	ADP
bracis-33593	190	23	the	the	DET
bracis-33593	190	24	second	second	ADJ
bracis-33593	190	25	episode	episode	NOUN
bracis-33593	190	26	.	.	PUNCT
bracis-33593	191	1	by	by	ADP
bracis-33593	191	2	dynamically	dynamically	ADV
bracis-33593	191	3	adjusting	adjust	VERB
bracis-33593	191	4	their	their	PRON
bracis-33593	191	5	routing	routing	NOUN
bracis-33593	191	6	network	network	NOUN
bracis-33593	191	7	,	,	PUNCT
bracis-33593	191	8	the	the	DET
bracis-33593	191	9	k	k	NOUN
bracis-33593	191	10	-	-	PUNCT
bracis-33593	191	11	models	model	NOUN
bracis-33593	191	12	demonstrate	demonstrate	VERB
bracis-33593	191	13	a	a	DET
bracis-33593	191	14	remarkable	remarkable	ADJ
bracis-33593	191	15	ability	ability	NOUN
bracis-33593	191	16	to	to	PART
bracis-33593	191	17	rapidly	rapidly	ADV
bracis-33593	191	18	adapt	adapt	VERB
bracis-33593	191	19	to	to	ADP
bracis-33593	191	20	the	the	DET
bracis-33593	191	21	new	new	ADJ
bracis-33593	191	22	environment	environment	NOUN
bracis-33593	191	23	,	,	PUNCT
bracis-33593	191	24	resulting	result	VERB
bracis-33593	191	25	in	in	ADP
bracis-33593	191	26	shorter	short	ADJ
bracis-33593	191	27	episode	episode	NOUN
bracis-33593	191	28	duration	duration	NOUN
bracis-33593	191	29	.	.	PUNCT
bracis-33593	192	1	fig	fig	NOUN
bracis-33593	192	2	.	.	PUNCT
bracis-33593	193	1	7	7	X
bracis-33593	193	2	.	.	X
bracis-33593	193	3	box	box	NOUN
bracis-33593	193	4	plot	plot	NOUN
bracis-33593	193	5	showing	show	VERB
bracis-33593	193	6	distribution	distribution	NOUN
bracis-33593	193	7	of	of	ADP
bracis-33593	193	8	steps	step	NOUN
bracis-33593	193	9	that	that	PRON
bracis-33593	193	10	each	each	DET
bracis-33593	193	11	model	model	NOUN
bracis-33593	193	12	spent	spend	VERB
bracis-33593	193	13	to	to	PART
bracis-33593	193	14	get	get	VERB
bracis-33593	193	15	to	to	ADP
bracis-33593	193	16	the	the	DET
bracis-33593	193	17	goal	goal	NOUN
bracis-33593	193	18	across	across	ADP
bracis-33593	193	19	the	the	DET
bracis-33593	193	20	second	second	ADJ
bracis-33593	193	21	episode	episode	NOUN
bracis-33593	193	22	of	of	ADP
bracis-33593	193	23	the	the	DET
bracis-33593	193	24	second	second	ADJ
bracis-33593	193	25	scenario	scenario	NOUN
bracis-33593	193	26	with	with	ADP
bracis-33593	193	27	blue	blue	ADJ
bracis-33593	193	28	being	be	AUX
bracis-33593	193	29	the	the	DET
bracis-33593	193	30	standard	standard	ADJ
bracis-33593	193	31	single	single	ADJ
bracis-33593	193	32	model	model	NOUN
bracis-33593	193	33	approach	approach	NOUN
bracis-33593	193	34	and	and	CCONJ
bracis-33593	193	35	orange	orange	NOUN
bracis-33593	193	36	being	be	AUX
bracis-33593	193	37	the	the	DET
bracis-33593	193	38	k	k	NOUN
bracis-33593	193	39	-	-	PUNCT
bracis-33593	193	40	model	model	NOUN
bracis-33593	193	41	segmented	segment	VERB
bracis-33593	193	42	approach	approach	NOUN
bracis-33593	193	43	(	(	PUNCT
bracis-33593	193	44	color	color	NOUN
bracis-33593	193	45	figure	figure	NOUN
bracis-33593	193	46	online	online	ADV
bracis-33593	193	47	)	)	PUNCT
bracis-33593	193	48	full	full	ADJ
bracis-33593	193	49	size	size	NOUN
bracis-33593	193	50	image	image	NOUN
bracis-33593	193	51	overall	overall	ADV
bracis-33593	193	52	,	,	PUNCT
bracis-33593	193	53	these	these	DET
bracis-33593	193	54	findings	finding	NOUN
bracis-33593	193	55	underscore	underscore	VERB
bracis-33593	193	56	the	the	DET
bracis-33593	193	57	effectiveness	effectiveness	NOUN
bracis-33593	193	58	of	of	ADP
bracis-33593	193	59	the	the	DET
bracis-33593	193	60	segmented	segmented	ADJ
bracis-33593	193	61	k	k	NOUN
bracis-33593	193	62	-	-	PUNCT
bracis-33593	193	63	models	model	NOUN
bracis-33593	193	64	approach	approach	NOUN
bracis-33593	193	65	in	in	ADP
bracis-33593	193	66	leveraging	leverage	VERB
bracis-33593	193	67	accumulated	accumulate	VERB
bracis-33593	193	68	knowledge	knowledge	NOUN
bracis-33593	193	69	from	from	ADP
bracis-33593	193	70	previous	previous	ADJ
bracis-33593	193	71	episodes	episode	NOUN
bracis-33593	193	72	to	to	PART
bracis-33593	193	73	facilitate	facilitate	VERB
bracis-33593	193	74	faster	fast	ADJ
bracis-33593	193	75	adaptation	adaptation	NOUN
bracis-33593	193	76	and	and	CCONJ
bracis-33593	193	77	improved	improved	ADJ
bracis-33593	193	78	performance	performance	NOUN
bracis-33593	193	79	in	in	ADP
bracis-33593	193	80	subsequent	subsequent	ADJ
bracis-33593	193	81	episodes	episode	NOUN
bracis-33593	193	82	.	.	PUNCT
bracis-33593	194	1	this	this	PRON
bracis-33593	194	2	highlights	highlight	VERB
bracis-33593	194	3	the	the	DET
bracis-33593	194	4	advantages	advantage	NOUN
bracis-33593	194	5	of	of	ADP
bracis-33593	194	6	employing	employ	VERB
bracis-33593	194	7	a	a	DET
bracis-33593	194	8	segmented	segment	VERB
bracis-33593	194	9	representation	representation	NOUN
bracis-33593	194	10	of	of	ADP
bracis-33593	194	11	the	the	DET
bracis-33593	194	12	world	world	NOUN
bracis-33593	194	13	’s	’s	PART
bracis-33593	194	14	dynamics	dynamic	NOUN
bracis-33593	194	15	,	,	PUNCT
bracis-33593	194	16	particularly	particularly	ADV
bracis-33593	194	17	in	in	ADP
bracis-33593	194	18	scenarios	scenario	NOUN
bracis-33593	194	19	where	where	SCONJ
bracis-33593	194	20	the	the	DET
bracis-33593	194	21	environment	environment	NOUN
bracis-33593	194	22	exhibits	exhibit	VERB
bracis-33593	194	23	dynamic	dynamic	ADJ
bracis-33593	194	24	changes	change	NOUN
bracis-33593	194	25	between	between	ADP
bracis-33593	194	26	episodes	episode	NOUN
bracis-33593	194	27	.	.	PUNCT
bracis-33593	195	1	7	7	NUM
bracis-33593	195	2	conclusion	conclusion	NOUN
bracis-33593	195	3	this	this	DET
bracis-33593	195	4	work	work	NOUN
bracis-33593	195	5	aims	aim	VERB
bracis-33593	195	6	to	to	PART
bracis-33593	195	7	demonstrate	demonstrate	VERB
bracis-33593	195	8	the	the	DET
bracis-33593	195	9	benefits	benefit	NOUN
bracis-33593	195	10	of	of	ADP
bracis-33593	195	11	leveraging	leverage	VERB
bracis-33593	195	12	segmented	segment	VERB
bracis-33593	195	13	dynamics	dynamic	NOUN
bracis-33593	195	14	in	in	ADP
bracis-33593	195	15	reinforcement	reinforcement	NOUN
bracis-33593	195	16	learning	learning	NOUN
bracis-33593	195	17	problems	problem	NOUN
bracis-33593	195	18	.	.	PUNCT
bracis-33593	196	1	by	by	ADP
bracis-33593	196	2	recognizing	recognize	VERB
bracis-33593	196	3	that	that	SCONJ
bracis-33593	196	4	the	the	DET
bracis-33593	196	5	environment	environment	NOUN
bracis-33593	196	6	’s	’s	PART
bracis-33593	196	7	dynamics	dynamic	NOUN
bracis-33593	196	8	can	can	AUX
bracis-33593	196	9	be	be	AUX
bracis-33593	196	10	effectively	effectively	ADV
bracis-33593	196	11	segmented	segment	VERB
bracis-33593	196	12	into	into	ADP
bracis-33593	196	13	distinct	distinct	ADJ
bracis-33593	196	14	regions	region	NOUN
bracis-33593	196	15	,	,	PUNCT
bracis-33593	196	16	we	we	PRON
bracis-33593	196	17	introduce	introduce	VERB
bracis-33593	196	18	a	a	DET
bracis-33593	196	19	novel	novel	ADJ
bracis-33593	196	20	approach	approach	NOUN
bracis-33593	196	21	that	that	PRON
bracis-33593	196	22	enhances	enhance	VERB
bracis-33593	196	23	the	the	DET
bracis-33593	196	24	efficiency	efficiency	NOUN
bracis-33593	196	25	of	of	ADP
bracis-33593	196	26	reinforcement	reinforcement	NOUN
bracis-33593	196	27	learning	learning	NOUN
bracis-33593	196	28	solutions	solution	NOUN
bracis-33593	196	29	.	.	PUNCT
bracis-33593	197	1	through	through	ADP
bracis-33593	197	2	the	the	DET
bracis-33593	197	3	utilization	utilization	NOUN
bracis-33593	197	4	of	of	ADP
bracis-33593	197	5	a	a	DET
bracis-33593	197	6	structured	structured	ADJ
bracis-33593	197	7	framework	framework	NOUN
bracis-33593	197	8	comprised	comprise	VERB
bracis-33593	197	9	of	of	ADP
bracis-33593	197	10	smaller	small	ADJ
bracis-33593	197	11	models	model	NOUN
bracis-33593	197	12	,	,	PUNCT
bracis-33593	197	13	the	the	DET
bracis-33593	197	14	agent	agent	NOUN
bracis-33593	197	15	gains	gain	VERB
bracis-33593	197	16	the	the	DET
bracis-33593	197	17	ability	ability	NOUN
bracis-33593	197	18	to	to	PART
bracis-33593	197	19	discern	discern	VERB
bracis-33593	197	20	situations	situation	NOUN
bracis-33593	197	21	where	where	SCONJ
bracis-33593	197	22	it	it	PRON
bracis-33593	197	23	can	can	AUX
bracis-33593	197	24	switch	switch	VERB
bracis-33593	197	25	between	between	ADP
bracis-33593	197	26	models	model	NOUN
bracis-33593	197	27	.	.	PUNCT
bracis-33593	198	1	this	this	DET
bracis-33593	198	2	segmentation	segmentation	NOUN
bracis-33593	198	3	allows	allow	VERB
bracis-33593	198	4	for	for	ADP
bracis-33593	198	5	improved	improved	ADJ
bracis-33593	198	6	generalization	generalization	NOUN
bracis-33593	198	7	across	across	ADP
bracis-33593	198	8	diverse	diverse	ADJ
bracis-33593	198	9	contexts	context	NOUN
bracis-33593	198	10	and	and	CCONJ
bracis-33593	198	11	facilitates	facilitate	VERB
bracis-33593	198	12	faster	fast	ADJ
bracis-33593	198	13	adaptation	adaptation	NOUN
bracis-33593	198	14	to	to	ADP
bracis-33593	198	15	familiar	familiar	ADJ
bracis-33593	198	16	dynamics	dynamic	NOUN
bracis-33593	198	17	within	within	ADP
bracis-33593	198	18	different	different	ADJ
bracis-33593	198	19	environments	environment	NOUN
bracis-33593	198	20	.	.	PUNCT
bracis-33593	199	1	by	by	ADP
bracis-33593	199	2	adopting	adopt	VERB
bracis-33593	199	3	this	this	DET
bracis-33593	199	4	approach	approach	NOUN
bracis-33593	199	5	,	,	PUNCT
bracis-33593	199	6	our	our	PRON
bracis-33593	199	7	research	research	NOUN
bracis-33593	199	8	highlights	highlight	VERB
bracis-33593	199	9	the	the	DET
bracis-33593	199	10	importance	importance	NOUN
bracis-33593	199	11	of	of	ADP
bracis-33593	199	12	understanding	understanding	NOUN
bracis-33593	199	13	and	and	CCONJ
bracis-33593	199	14	leveraging	leverage	VERB
bracis-33593	199	15	the	the	DET
bracis-33593	199	16	underlying	underlie	VERB
bracis-33593	199	17	structure	structure	NOUN
bracis-33593	199	18	of	of	ADP
bracis-33593	199	19	the	the	DET
bracis-33593	199	20	environment	environment	NOUN
bracis-33593	199	21	’s	’s	PART
bracis-33593	199	22	dynamics	dynamic	NOUN
bracis-33593	199	23	.	.	PUNCT
bracis-33593	200	1	by	by	ADP
bracis-33593	200	2	incorporating	incorporate	VERB
bracis-33593	200	3	segmented	segment	VERB
bracis-33593	200	4	models	model	NOUN
bracis-33593	200	5	into	into	ADP
bracis-33593	200	6	the	the	DET
bracis-33593	200	7	learning	learning	NOUN
bracis-33593	200	8	process	process	NOUN
bracis-33593	200	9	,	,	PUNCT
bracis-33593	200	10	we	we	PRON
bracis-33593	200	11	empower	empower	VERB
bracis-33593	200	12	agents	agent	NOUN
bracis-33593	200	13	to	to	PART
bracis-33593	200	14	navigate	navigate	VERB
bracis-33593	200	15	complex	complex	ADJ
bracis-33593	200	16	environments	environment	NOUN
bracis-33593	200	17	more	more	ADV
bracis-33593	200	18	effectively	effectively	ADV
bracis-33593	200	19	,	,	PUNCT
bracis-33593	200	20	ultimately	ultimately	ADV
bracis-33593	200	21	leading	lead	VERB
bracis-33593	200	22	to	to	ADP
bracis-33593	200	23	enhanced	enhanced	ADJ
bracis-33593	200	24	performance	performance	NOUN
bracis-33593	200	25	and	and	CCONJ
bracis-33593	200	26	adaptability	adaptability	NOUN
bracis-33593	200	27	in	in	ADP
bracis-33593	200	28	a	a	DET
bracis-33593	200	29	variety	variety	NOUN
bracis-33593	200	30	of	of	ADP
bracis-33593	200	31	reinforcement	reinforcement	NOUN
bracis-33593	200	32	learning	learning	NOUN
bracis-33593	200	33	tasks	task	NOUN
bracis-33593	200	34	.	.	PUNCT
bracis-33593	201	1	one	one	NUM
bracis-33593	201	2	limitation	limitation	NOUN
bracis-33593	201	3	of	of	ADP
bracis-33593	201	4	our	our	PRON
bracis-33593	201	5	approach	approach	NOUN
bracis-33593	201	6	is	be	AUX
bracis-33593	201	7	the	the	DET
bracis-33593	201	8	requirement	requirement	NOUN
bracis-33593	201	9	to	to	PART
bracis-33593	201	10	specify	specify	VERB
bracis-33593	201	11	the	the	DET
bracis-33593	201	12	number	number	NOUN
bracis-33593	201	13	of	of	ADP
bracis-33593	201	14	models	model	NOUN
bracis-33593	201	15	as	as	ADP
bracis-33593	201	16	input	input	NOUN
bracis-33593	201	17	,	,	PUNCT
bracis-33593	201	18	which	which	PRON
bracis-33593	201	19	may	may	AUX
bracis-33593	201	20	be	be	AUX
bracis-33593	201	21	unknown	unknown	ADJ
bracis-33593	201	22	in	in	ADP
bracis-33593	201	23	practical	practical	ADJ
bracis-33593	201	24	applications	application	NOUN
bracis-33593	201	25	.	.	PUNCT
bracis-33593	202	1	this	this	PRON
bracis-33593	202	2	presents	present	VERB
bracis-33593	202	3	a	a	DET
bracis-33593	202	4	challenge	challenge	NOUN
bracis-33593	202	5	,	,	PUNCT
bracis-33593	202	6	as	as	ADP
bracis-33593	202	7	determining	determine	VERB
bracis-33593	202	8	the	the	DET
bracis-33593	202	9	optimal	optimal	ADJ
bracis-33593	202	10	number	number	NOUN
bracis-33593	202	11	of	of	ADP
bracis-33593	202	12	models	model	NOUN
bracis-33593	202	13	can	can	AUX
bracis-33593	202	14	be	be	AUX
bracis-33593	202	15	non	non	ADJ
bracis-33593	202	16	-	-	ADJ
bracis-33593	202	17	trivial	trivial	ADJ
bracis-33593	202	18	and	and	CCONJ
bracis-33593	202	19	may	may	AUX
bracis-33593	202	20	require	require	VERB
bracis-33593	202	21	prior	prior	ADJ
bracis-33593	202	22	knowledge	knowledge	NOUN
bracis-33593	202	23	or	or	CCONJ
bracis-33593	202	24	experimentation	experimentation	NOUN
bracis-33593	202	25	.	.	PUNCT
bracis-33593	203	1	to	to	PART
bracis-33593	203	2	address	address	VERB
bracis-33593	203	3	this	this	DET
bracis-33593	203	4	limitation	limitation	NOUN
bracis-33593	203	5	,	,	PUNCT
bracis-33593	203	6	future	future	ADJ
bracis-33593	203	7	research	research	NOUN
bracis-33593	203	8	could	could	AUX
bracis-33593	203	9	explore	explore	VERB
bracis-33593	203	10	methods	method	NOUN
bracis-33593	203	11	for	for	ADP
bracis-33593	203	12	automatically	automatically	ADV
bracis-33593	203	13	determining	determine	VERB
bracis-33593	203	14	the	the	DET
bracis-33593	203	15	appropriate	appropriate	ADJ
bracis-33593	203	16	number	number	NOUN
bracis-33593	203	17	of	of	ADP
bracis-33593	203	18	models	model	NOUN
bracis-33593	203	19	based	base	VERB
bracis-33593	203	20	on	on	ADP
bracis-33593	203	21	the	the	DET
bracis-33593	203	22	characteristics	characteristic	NOUN
bracis-33593	203	23	of	of	ADP
bracis-33593	203	24	the	the	DET
bracis-33593	203	25	environment	environment	NOUN
bracis-33593	203	26	or	or	CCONJ
bracis-33593	203	27	through	through	ADP
bracis-33593	203	28	data	data	NOUN
bracis-33593	203	29	-	-	PUNCT
bracis-33593	203	30	driven	drive	VERB
bracis-33593	203	31	approaches	approach	NOUN
bracis-33593	203	32	.	.	PUNCT
bracis-33593	204	1	techniques	technique	NOUN
bracis-33593	204	2	such	such	ADJ
bracis-33593	204	3	as	as	ADP
bracis-33593	204	4	model	model	NOUN
bracis-33593	204	5	selection	selection	NOUN
bracis-33593	204	6	criteria	criterion	NOUN
bracis-33593	204	7	,	,	PUNCT
bracis-33593	204	8	bayesian	bayesian	NOUN
bracis-33593	204	9	optimization	optimization	NOUN
bracis-33593	204	10	,	,	PUNCT
bracis-33593	204	11	or	or	CCONJ
bracis-33593	204	12	clustering	cluster	VERB
bracis-33593	204	13	algorithms	algorithm	NOUN
bracis-33593	204	14	could	could	AUX
bracis-33593	204	15	be	be	AUX
bracis-33593	204	16	employed	employ	VERB
bracis-33593	204	17	to	to	PART
bracis-33593	204	18	dynamically	dynamically	ADV
bracis-33593	204	19	adjust	adjust	VERB
bracis-33593	204	20	the	the	DET
bracis-33593	204	21	number	number	NOUN
bracis-33593	204	22	of	of	ADP
bracis-33593	204	23	models	model	NOUN
bracis-33593	204	24	based	base	VERB
bracis-33593	204	25	on	on	ADP
bracis-33593	204	26	observed	observed	ADJ
bracis-33593	204	27	data	datum	NOUN
bracis-33593	204	28	and	and	CCONJ
bracis-33593	204	29	performance	performance	NOUN
bracis-33593	204	30	metrics	metric	NOUN
bracis-33593	204	31	.	.	PUNCT
bracis-33593	205	1	additionally	additionally	ADV
bracis-33593	205	2	,	,	PUNCT
bracis-33593	205	3	the	the	DET
bracis-33593	205	4	flexibility	flexibility	NOUN
bracis-33593	205	5	of	of	ADP
bracis-33593	205	6	our	our	PRON
bracis-33593	205	7	approach	approach	NOUN
bracis-33593	205	8	allows	allow	VERB
bracis-33593	205	9	for	for	ADP
bracis-33593	205	10	the	the	DET
bracis-33593	205	11	incorporation	incorporation	NOUN
bracis-33593	205	12	of	of	ADP
bracis-33593	205	13	adaptive	adaptive	ADJ
bracis-33593	205	14	mechanisms	mechanism	NOUN
bracis-33593	205	15	that	that	PRON
bracis-33593	205	16	can	can	AUX
bracis-33593	205	17	dynamically	dynamically	ADV
bracis-33593	205	18	adjust	adjust	VERB
bracis-33593	205	19	the	the	DET
bracis-33593	205	20	number	number	NOUN
bracis-33593	205	21	of	of	ADP
bracis-33593	205	22	models	model	NOUN
bracis-33593	205	23	based	base	VERB
bracis-33593	205	24	on	on	ADP
bracis-33593	205	25	feedback	feedback	NOUN
bracis-33593	205	26	from	from	ADP
bracis-33593	205	27	the	the	DET
bracis-33593	205	28	learning	learning	NOUN
bracis-33593	205	29	process	process	NOUN
bracis-33593	205	30	.	.	PUNCT
bracis-33593	206	1	by	by	ADP
bracis-33593	206	2	incorporating	incorporate	VERB
bracis-33593	206	3	such	such	ADJ
bracis-33593	206	4	adaptive	adaptive	ADJ
bracis-33593	206	5	mechanisms	mechanism	NOUN
bracis-33593	206	6	,	,	PUNCT
bracis-33593	206	7	our	our	PRON
bracis-33593	206	8	approach	approach	NOUN
bracis-33593	206	9	can	can	AUX
bracis-33593	206	10	become	become	VERB
bracis-33593	206	11	more	more	ADV
bracis-33593	206	12	robust	robust	ADJ
bracis-33593	206	13	and	and	CCONJ
bracis-33593	206	14	applicable	applicable	ADJ
bracis-33593	206	15	to	to	ADP
bracis-33593	206	16	a	a	DET
bracis-33593	206	17	wider	wide	ADJ
bracis-33593	206	18	range	range	NOUN
bracis-33593	206	19	of	of	ADP
bracis-33593	206	20	real	real	ADJ
bracis-33593	206	21	-	-	PUNCT
bracis-33593	206	22	world	world	NOUN
bracis-33593	206	23	scenarios	scenario	NOUN
bracis-33593	206	24	where	where	SCONJ
bracis-33593	206	25	the	the	DET
bracis-33593	206	26	number	number	NOUN
bracis-33593	206	27	of	of	ADP
bracis-33593	206	28	models	model	NOUN
bracis-33593	206	29	may	may	AUX
bracis-33593	206	30	vary	vary	VERB
bracis-33593	206	31	or	or	CCONJ
bracis-33593	206	32	change	change	VERB
bracis-33593	206	33	over	over	ADP
bracis-33593	206	34	time	time	NOUN
bracis-33593	206	35	.	.	PUNCT
bracis-33593	207	1	overall	overall	ADV
bracis-33593	207	2	,	,	PUNCT
bracis-33593	207	3	addressing	address	VERB
bracis-33593	207	4	the	the	DET
bracis-33593	207	5	limitation	limitation	NOUN
bracis-33593	207	6	of	of	ADP
bracis-33593	207	7	requiring	require	VERB
bracis-33593	207	8	a	a	DET
bracis-33593	207	9	predefined	predefine	VERB
bracis-33593	207	10	number	number	NOUN
bracis-33593	207	11	of	of	ADP
bracis-33593	207	12	models	model	NOUN
bracis-33593	207	13	is	be	AUX
bracis-33593	207	14	an	an	DET
bracis-33593	207	15	important	important	ADJ
bracis-33593	207	16	direction	direction	NOUN
bracis-33593	207	17	for	for	ADP
bracis-33593	207	18	future	future	ADJ
bracis-33593	207	19	research	research	NOUN
bracis-33593	207	20	in	in	ADP
bracis-33593	207	21	order	order	NOUN
bracis-33593	207	22	to	to	PART
bracis-33593	207	23	enhance	enhance	VERB
bracis-33593	207	24	the	the	DET
bracis-33593	207	25	scalability	scalability	NOUN
bracis-33593	207	26	and	and	CCONJ
bracis-33593	207	27	applicability	applicability	NOUN
bracis-33593	207	28	of	of	ADP
bracis-33593	207	29	our	our	PRON
bracis-33593	207	30	approach	approach	NOUN
bracis-33593	207	31	in	in	ADP
bracis-33593	207	32	practical	practical	ADJ
bracis-33593	207	33	reinforcement	reinforcement	NOUN
bracis-33593	207	34	learning	learning	NOUN
bracis-33593	207	35	applications	application	NOUN
bracis-33593	207	36	.	.	PUNCT
bracis-33593	208	1	references	reference	NOUN
bracis-33593	208	2	alegre	alegre	VERB
bracis-33593	208	3	,	,	PUNCT
bracis-33593	208	4	l.n	l.n	PROPN
bracis-33593	208	5	.	.	PROPN
bracis-33593	208	6	,	,	PUNCT
bracis-33593	208	7	bazzan	bazzan	NOUN
bracis-33593	208	8	,	,	PUNCT
bracis-33593	208	9	a.l.c	a.l.c	NOUN
bracis-33593	208	10	.	.	PUNCT
bracis-33593	208	11	,	,	PUNCT
bracis-33593	208	12	da	da	PROPN
bracis-33593	208	13	 	 	SPACE
bracis-33593	208	14	silva	silva	PROPN
bracis-33593	208	15	,	,	PUNCT
bracis-33593	208	16	b.c	b.c	PROPN
bracis-33593	208	17	.	.	PROPN
bracis-33593	208	18	:	:	PUNCT
bracis-33593	209	1	minimum	minimum	ADJ
bracis-33593	209	2	-	-	PUNCT
bracis-33593	209	3	delay	delay	NOUN
bracis-33593	209	4	adaptation	adaptation	NOUN
bracis-33593	209	5	in	in	ADP
bracis-33593	209	6	non	non	ADJ
bracis-33593	209	7	-	-	ADJ
bracis-33593	209	8	stationary	stationary	ADJ
bracis-33593	209	9	reinforcement	reinforcement	NOUN
bracis-33593	209	10	learning	learning	NOUN
bracis-33593	209	11	via	via	ADP
bracis-33593	209	12	online	online	ADJ
bracis-33593	209	13	high	high	ADJ
bracis-33593	209	14	-	-	PUNCT
bracis-33593	209	15	confidence	confidence	NOUN
bracis-33593	209	16	change	change	NOUN
bracis-33593	209	17	-	-	PUNCT
bracis-33593	209	18	point	point	NOUN
bracis-33593	209	19	detection	detection	NOUN
bracis-33593	209	20	.	.	PUNCT
bracis-33593	210	1	in	in	ADP
bracis-33593	210	2	:	:	PUNCT
bracis-33593	210	3	proceedings	proceeding	NOUN
bracis-33593	210	4	of	of	ADP
bracis-33593	210	5	the	the	DET
bracis-33593	210	6	20th	20th	ADJ
bracis-33593	210	7	international	international	ADJ
bracis-33593	210	8	conference	conference	NOUN
bracis-33593	210	9	on	on	ADP
bracis-33593	210	10	autonomous	autonomous	ADJ
bracis-33593	210	11	agents	agent	NOUN
bracis-33593	210	12	and	and	CCONJ
bracis-33593	210	13	multiagent	multiagent	NOUN
bracis-33593	210	14	systems	system	NOUN
bracis-33593	210	15	,	,	PUNCT
bracis-33593	210	16	aamas	aamas	PROPN
bracis-33593	210	17	’	’	PROPN
bracis-33593	210	18	21	21	NUM
bracis-33593	210	19	,	,	PUNCT
bracis-33593	210	20	pp	pp	ADJ
bracis-33593	210	21	.	.	PUNCT
bracis-33593	211	1	97–105	97–105	NUM
bracis-33593	211	2	.	.	PUNCT
bracis-33593	212	1	international	international	PROPN
bracis-33593	212	2	foundation	foundation	PROPN
bracis-33593	212	3	for	for	ADP
bracis-33593	212	4	autonomous	autonomous	ADJ
bracis-33593	212	5	agents	agent	NOUN
bracis-33593	212	6	and	and	CCONJ
bracis-33593	212	7	multiagent	multiagent	NOUN
bracis-33593	212	8	systems	system	NOUN
bracis-33593	212	9	,	,	PUNCT
bracis-33593	212	10	richland	richland	PROPN
bracis-33593	212	11	(	(	PUNCT
bracis-33593	212	12	2021	2021	NUM
bracis-33593	212	13	)	)	PUNCT
bracis-33593	212	14	google	google	NOUN
bracis-33593	212	15	scholar	scholar	NOUN
bracis-33593	212	16	  	  	SPACE
bracis-33593	212	17	chitnis	chitnis	PROPN
bracis-33593	212	18	,	,	PUNCT
bracis-33593	212	19	r.	r.	PROPN
bracis-33593	212	20	,	,	PUNCT
bracis-33593	212	21	silver	silver	NOUN
bracis-33593	212	22	,	,	PUNCT
bracis-33593	212	23	t.	t.	PROPN
bracis-33593	212	24	,	,	PUNCT
bracis-33593	212	25	tenenbaum	tenenbaum	PROPN
bracis-33593	212	26	,	,	PUNCT
bracis-33593	212	27	j.	j.	PROPN
bracis-33593	212	28	,	,	PUNCT
bracis-33593	212	29	kaelbling	kaelble	VERB
bracis-33593	212	30	,	,	PUNCT
bracis-33593	212	31	l.p	l.p	PROPN
bracis-33593	212	32	.	.	PROPN
bracis-33593	212	33	,	,	PUNCT
bracis-33593	212	34	lozano	lozano	PROPN
bracis-33593	212	35	-	-	PUNCT
bracis-33593	212	36	pérez	pérez	NOUN
bracis-33593	212	37	,	,	PUNCT
bracis-33593	212	38	t.	t.	NOUN
bracis-33593	212	39	:	:	PUNCT
bracis-33593	212	40	glib	glib	ADJ
bracis-33593	212	41	:	:	PUNCT
bracis-33593	212	42	efficient	efficient	ADJ
bracis-33593	212	43	exploration	exploration	NOUN
bracis-33593	212	44	for	for	ADP
bracis-33593	212	45	relational	relational	ADJ
bracis-33593	212	46	model	model	NOUN
bracis-33593	212	47	-	-	PUNCT
bracis-33593	212	48	based	base	VERB
bracis-33593	212	49	reinforcement	reinforcement	NOUN
bracis-33593	212	50	learning	learning	NOUN
bracis-33593	212	51	via	via	ADP
bracis-33593	212	52	goal	goal	NOUN
bracis-33593	212	53	-	-	PUNCT
bracis-33593	212	54	literal	literal	ADJ
bracis-33593	212	55	babbling	babbling	NOUN
bracis-33593	212	56	,	,	PUNCT
bracis-33593	212	57	vol	vol	NOUN
bracis-33593	212	58	.	.	PUNCT
bracis-33593	212	59	 	 	SPACE
bracis-33593	212	60	13b	13b	NOUN
bracis-33593	212	61	,	,	PUNCT
bracis-33593	212	62	pp	pp	ADV
bracis-33593	212	63	.	.	PUNCT
bracis-33593	213	1	11782–11791	11782–11791	NUM
bracis-33593	213	2	(	(	PUNCT
bracis-33593	213	3	2021	2021	NUM
bracis-33593	213	4	)	)	PUNCT
bracis-33593	213	5	.	.	PUNCT
bracis-33593	214	1	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85100285730&partnerid=40	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85100285730&partnerid=40	PROPN
bracis-33593	214	2	&	&	CCONJ
bracis-33593	214	3	md5=4e2e2fca957818475950cdcd634527fe	md5=4e2e2fca957818475950cdcd634527fe	PROPN
bracis-33593	214	4	doya	doya	PROPN
bracis-33593	214	5	,	,	PUNCT
bracis-33593	214	6	k.	k.	PROPN
bracis-33593	214	7	,	,	PUNCT
bracis-33593	214	8	samejima	samejima	PROPN
bracis-33593	214	9	,	,	PUNCT
bracis-33593	214	10	k.	k.	PROPN
bracis-33593	214	11	,	,	PUNCT
bracis-33593	214	12	katagiri	katagiri	PROPN
bracis-33593	214	13	,	,	PUNCT
bracis-33593	214	14	k.i	k.i	PROPN
bracis-33593	214	15	.	.	PROPN
bracis-33593	214	16	,	,	PUNCT
bracis-33593	214	17	kawato	kawato	PROPN
bracis-33593	214	18	,	,	PUNCT
bracis-33593	214	19	m.	m.	NOUN
bracis-33593	214	20	:	:	PUNCT
bracis-33593	214	21	multiple	multiple	ADJ
bracis-33593	214	22	model	model	NOUN
bracis-33593	214	23	-	-	PUNCT
bracis-33593	214	24	based	base	VERB
bracis-33593	214	25	reinforcement	reinforcement	NOUN
bracis-33593	214	26	learning	learning	NOUN
bracis-33593	214	27	.	.	PUNCT
bracis-33593	215	1	neural	neural	ADJ
bracis-33593	215	2	comput	comput	NOUN
bracis-33593	215	3	.	.	PUNCT
bracis-33593	216	1	14(6	14(6	NOUN
bracis-33593	216	2	)	)	PUNCT
bracis-33593	216	3	,	,	PUNCT
bracis-33593	216	4	1347–1369	1347–1369	NUM
bracis-33593	216	5	(	(	PUNCT
bracis-33593	216	6	2002	2002	NUM
bracis-33593	216	7	)	)	PUNCT
bracis-33593	216	8	.	.	PUNCT
bracis-33593	217	1	https://doi.org/10.1162/089976602753712972	https://doi.org/10.1162/089976602753712972	PROPN
bracis-33593	217	2	fox	fox	PROPN
bracis-33593	217	3	,	,	PUNCT
bracis-33593	217	4	r.	r.	PROPN
bracis-33593	217	5	,	,	PUNCT
bracis-33593	217	6	krishnan	krishnan	PROPN
bracis-33593	217	7	,	,	PUNCT
bracis-33593	217	8	s.	s.	PROPN
bracis-33593	217	9	,	,	PUNCT
bracis-33593	217	10	stoica	stoica	PROPN
bracis-33593	217	11	,	,	PUNCT
bracis-33593	217	12	i.	i.	PROPN
bracis-33593	217	13	,	,	PUNCT
bracis-33593	217	14	goldberg	goldberg	PROPN
bracis-33593	217	15	,	,	PUNCT
bracis-33593	217	16	k.	k.	PROPN
bracis-33593	217	17	:	:	PUNCT
bracis-33593	217	18	multi	multi	ADJ
bracis-33593	217	19	-	-	ADJ
bracis-33593	217	20	level	level	ADJ
bracis-33593	217	21	discovery	discovery	NOUN
bracis-33593	217	22	of	of	ADP
bracis-33593	217	23	deep	deep	ADJ
bracis-33593	217	24	options	option	NOUN
bracis-33593	217	25	(	(	PUNCT
bracis-33593	217	26	2017	2017	NUM
bracis-33593	217	27	)	)	PUNCT
bracis-33593	217	28	google	google	PROPN
bracis-33593	217	29	scholar	scholar	NOUN
bracis-33593	217	30	  	  	SPACE
bracis-33593	217	31	hallak	hallak	PROPN
bracis-33593	217	32	,	,	PUNCT
bracis-33593	217	33	a.	a.	PROPN
bracis-33593	217	34	,	,	PUNCT
bracis-33593	217	35	castro	castro	PROPN
bracis-33593	217	36	,	,	PUNCT
bracis-33593	217	37	d.	d.	PROPN
bracis-33593	217	38	,	,	PUNCT
bracis-33593	217	39	mannor	mannor	PROPN
bracis-33593	217	40	,	,	PUNCT
bracis-33593	217	41	s.	s.	PROPN
bracis-33593	217	42	:	:	PUNCT
bracis-33593	217	43	contextual	contextual	ADJ
bracis-33593	217	44	markov	markov	NOUN
bracis-33593	217	45	decision	decision	NOUN
bracis-33593	217	46	processes	process	NOUN
bracis-33593	217	47	(	(	PUNCT
bracis-33593	217	48	2015	2015	NUM
bracis-33593	217	49	)	)	PUNCT
bracis-33593	217	50	google	google	PROPN
bracis-33593	217	51	scholar	scholar	NOUN
bracis-33593	217	52	  	  	SPACE
bracis-33593	217	53	kidambi	kidambi	PROPN
bracis-33593	217	54	,	,	PUNCT
bracis-33593	217	55	r.	r.	PROPN
bracis-33593	217	56	,	,	PUNCT
bracis-33593	217	57	rajeswaran	rajeswaran	VERB
bracis-33593	217	58	,	,	PUNCT
bracis-33593	217	59	a.	a.	NOUN
bracis-33593	217	60	,	,	PUNCT
bracis-33593	217	61	netrapalli	netrapalli	ADV
bracis-33593	217	62	,	,	PUNCT
bracis-33593	217	63	p.	p.	NOUN
bracis-33593	217	64	,	,	PUNCT
bracis-33593	217	65	joachims	joachim	NOUN
bracis-33593	217	66	,	,	PUNCT
bracis-33593	217	67	t.	t.	PROPN
bracis-33593	217	68	:	:	PUNCT
bracis-33593	217	69	morel	morel	PROPN
bracis-33593	217	70	:	:	PUNCT
bracis-33593	217	71	model	model	NOUN
bracis-33593	217	72	-	-	PUNCT
bracis-33593	217	73	based	base	VERB
bracis-33593	217	74	offline	offline	ADJ
bracis-33593	217	75	reinforcement	reinforcement	NOUN
bracis-33593	217	76	learning	learning	NOUN
bracis-33593	217	77	.	.	PUNCT
bracis-33593	218	1	in	in	ADP
bracis-33593	218	2	:	:	PUNCT
bracis-33593	218	3	nips	nip	NOUN
bracis-33593	218	4	2020	2020	NUM
bracis-33593	218	5	.	.	PUNCT
bracis-33593	219	1	curran	curran	PROPN
bracis-33593	219	2	associates	associates	PROPN
bracis-33593	219	3	inc	inc	PROPN
bracis-33593	219	4	.	.	PROPN
bracis-33593	219	5	,	,	PUNCT
bracis-33593	219	6	red	red	ADJ
bracis-33593	219	7	hook	hook	NOUN
bracis-33593	219	8	(	(	PUNCT
bracis-33593	219	9	2020	2020	NUM
bracis-33593	219	10	)	)	PUNCT
bracis-33593	219	11	google	google	PROPN
bracis-33593	219	12	scholar	scholar	NOUN
bracis-33593	219	13	  	  	SPACE
bracis-33593	219	14	li	li	PROPN
bracis-33593	219	15	,	,	PUNCT
bracis-33593	219	16	j.	j.	PROPN
bracis-33593	219	17	,	,	PUNCT
bracis-33593	219	18	tang	tang	PROPN
bracis-33593	219	19	,	,	PUNCT
bracis-33593	219	20	c.	c.	PROPN
bracis-33593	219	21	,	,	PUNCT
bracis-33593	219	22	tomizuka	tomizuka	PROPN
bracis-33593	219	23	,	,	PUNCT
bracis-33593	219	24	m.	m.	NOUN
bracis-33593	219	25	,	,	PUNCT
bracis-33593	219	26	zhan	zhan	PROPN
bracis-33593	219	27	,	,	PUNCT
bracis-33593	219	28	w.	w.	PROPN
bracis-33593	219	29	:	:	PUNCT
bracis-33593	219	30	hierarchical	hierarchical	ADJ
bracis-33593	219	31	planning	planning	NOUN
bracis-33593	219	32	through	through	ADP
bracis-33593	219	33	goal	goal	NOUN
bracis-33593	219	34	-	-	PUNCT
bracis-33593	219	35	conditioned	condition	VERB
bracis-33593	219	36	offline	offline	ADJ
bracis-33593	219	37	reinforcement	reinforcement	NOUN
bracis-33593	219	38	learning	learning	NOUN
bracis-33593	219	39	.	.	PUNCT
bracis-33593	220	1	ieee	ieee	PROPN
bracis-33593	220	2	rob	rob	PROPN
bracis-33593	220	3	.	.	PUNCT
bracis-33593	221	1	autom	autom	PROPN
bracis-33593	221	2	.	.	PUNCT
bracis-33593	222	1	lett	lett	PROPN
bracis-33593	222	2	.	.	PUNCT
bracis-33593	223	1	7(4	7(4	NUM
bracis-33593	223	2	)	)	PUNCT
bracis-33593	223	3	,	,	PUNCT
bracis-33593	223	4	10216–10223	10216–10223	NUM
bracis-33593	223	5	(	(	PUNCT
bracis-33593	223	6	2022	2022	NUM
bracis-33593	223	7	)	)	PUNCT
bracis-33593	223	8	.	.	PUNCT
bracis-33593	224	1	https://doi.org/10.1109/lra.2022.3190100	https://doi.org/10.1109/lra.2022.3190100	PROPN
bracis-33593	224	2	article	article	NOUN
bracis-33593	224	3	  	  	SPACE
bracis-33593	224	4	math	math	NOUN
bracis-33593	224	5	  	  	SPACE
bracis-33593	224	6	google	google	PROPN
bracis-33593	224	7	scholar	scholar	NOUN
bracis-33593	224	8	  	  	SPACE
bracis-33593	224	9	li	li	PROPN
bracis-33593	224	10	,	,	PUNCT
bracis-33593	224	11	t.	t.	PROPN
bracis-33593	224	12	,	,	PUNCT
bracis-33593	224	13	lambert	lambert	PROPN
bracis-33593	224	14	,	,	PUNCT
bracis-33593	224	15	n.	n.	PROPN
bracis-33593	224	16	,	,	PUNCT
bracis-33593	224	17	calandra	calandra	PROPN
bracis-33593	224	18	,	,	PUNCT
bracis-33593	224	19	r.	r.	PROPN
bracis-33593	224	20	,	,	PUNCT
bracis-33593	224	21	meier	meier	PROPN
bracis-33593	224	22	,	,	PUNCT
bracis-33593	224	23	f.	f.	PROPN
bracis-33593	224	24	,	,	PUNCT
bracis-33593	224	25	rai	rai	PROPN
bracis-33593	224	26	,	,	PUNCT
bracis-33593	224	27	a.	a.	NOUN
bracis-33593	224	28	:	:	PUNCT
bracis-33593	224	29	learning	learn	VERB
bracis-33593	224	30	generalizable	generalizable	ADJ
bracis-33593	224	31	locomotion	locomotion	NOUN
bracis-33593	224	32	skills	skill	NOUN
bracis-33593	224	33	with	with	ADP
bracis-33593	224	34	hierarchical	hierarchical	ADJ
bracis-33593	224	35	reinforcement	reinforcement	NOUN
bracis-33593	224	36	learning	learning	NOUN
bracis-33593	224	37	.	.	PUNCT
bracis-33593	225	1	in	in	ADP
bracis-33593	225	2	:	:	PUNCT
bracis-33593	225	3	2020	2020	NUM
bracis-33593	225	4	ieee	ieee	PROPN
bracis-33593	225	5	international	international	ADJ
bracis-33593	225	6	conference	conference	NOUN
bracis-33593	225	7	on	on	ADP
bracis-33593	225	8	robotics	robotic	NOUN
bracis-33593	225	9	and	and	CCONJ
bracis-33593	225	10	automation	automation	NOUN
bracis-33593	225	11	(	(	PUNCT
bracis-33593	225	12	icra	icra	PROPN
bracis-33593	225	13	)	)	PUNCT
bracis-33593	225	14	,	,	PUNCT
bracis-33593	225	15	pp	pp	ADP
bracis-33593	225	16	.	.	PUNCT
bracis-33593	226	1	413–419	413–419	NUM
bracis-33593	226	2	(	(	PUNCT
bracis-33593	226	3	2020	2020	NUM
bracis-33593	226	4	)	)	PUNCT
bracis-33593	226	5	.	.	PUNCT
bracis-33593	227	1	https://doi.org/10.1109/icra40945.2020.9196642	https://doi.org/10.1109/icra40945.2020.9196642	PROPN
bracis-33593	227	2	lin	lin	PROPN
bracis-33593	227	3	,	,	PUNCT
bracis-33593	227	4	d.	d.	PROPN
bracis-33593	227	5	,	,	PUNCT
bracis-33593	227	6	zhang	zhang	PROPN
bracis-33593	227	7	,	,	PUNCT
bracis-33593	227	8	j.	j.	PROPN
bracis-33593	227	9	:	:	PUNCT
bracis-33593	227	10	a	a	DET
bracis-33593	227	11	reinforcement	reinforcement	NOUN
bracis-33593	227	12	learning	learning	NOUN
bracis-33593	227	13	based	base	VERB
bracis-33593	227	14	rrt	rrt	NOUN
bracis-33593	227	15	algorithm	algorithm	NOUN
bracis-33593	227	16	with	with	ADP
bracis-33593	227	17	value	value	NOUN
bracis-33593	227	18	estimation	estimation	NOUN
bracis-33593	227	19	.	.	PUNCT
bracis-33593	228	1	in	in	ADP
bracis-33593	228	2	:	:	PUNCT
bracis-33593	228	3	2022	2022	NUM
bracis-33593	228	4	3rd	3rd	ADJ
bracis-33593	228	5	international	international	ADJ
bracis-33593	228	6	conference	conference	NOUN
bracis-33593	228	7	on	on	ADP
bracis-33593	228	8	big	big	ADJ
bracis-33593	228	9	data	datum	NOUN
bracis-33593	228	10	,	,	PUNCT
bracis-33593	228	11	artificial	artificial	ADJ
bracis-33593	228	12	intelligence	intelligence	NOUN
bracis-33593	228	13	and	and	CCONJ
bracis-33593	228	14	internet	internet	NOUN
bracis-33593	228	15	of	of	ADP
bracis-33593	228	16	things	thing	NOUN
bracis-33593	228	17	engineering	engineering	NOUN
bracis-33593	228	18	(	(	PUNCT
bracis-33593	228	19	icbaie	icbaie	PROPN
bracis-33593	228	20	)	)	PUNCT
bracis-33593	228	21	,	,	PUNCT
bracis-33593	228	22	pp	pp	ADP
bracis-33593	228	23	.	.	PUNCT
bracis-33593	229	1	619–624	619–624	NUM
bracis-33593	229	2	(	(	PUNCT
bracis-33593	229	3	2022	2022	NUM
bracis-33593	229	4	)	)	PUNCT
bracis-33593	229	5	.	.	PUNCT
bracis-33593	230	1	https://doi.org/10.1109/icbaie56435.2022.9985840	https://doi.org/10.1109/icbaie56435.2022.9985840	PROPN
bracis-33593	230	2	ng	ng	PROPN
bracis-33593	230	3	,	,	PUNCT
bracis-33593	230	4	j.h.a	j.h.a	NOUN
bracis-33593	230	5	.	.	PROPN
bracis-33593	230	6	,	,	PUNCT
bracis-33593	230	7	petrick	petrick	PROPN
bracis-33593	230	8	,	,	PUNCT
bracis-33593	230	9	r.p.a	r.p.a	NOUN
bracis-33593	230	10	.	.	PUNCT
bracis-33593	230	11	:	:	PUNCT
bracis-33593	231	1	incremental	incremental	ADJ
bracis-33593	231	2	learning	learning	NOUN
bracis-33593	231	3	of	of	ADP
bracis-33593	231	4	planning	planning	NOUN
bracis-33593	231	5	actions	action	NOUN
bracis-33593	231	6	in	in	ADP
bracis-33593	231	7	model	model	NOUN
bracis-33593	231	8	-	-	PUNCT
bracis-33593	231	9	based	base	VERB
bracis-33593	231	10	reinforcement	reinforcement	NOUN
bracis-33593	231	11	learning	learning	NOUN
bracis-33593	231	12	.	.	PUNCT
bracis-33593	232	1	in	in	ADP
bracis-33593	232	2	:	:	PUNCT
bracis-33593	232	3	proceedings	proceeding	NOUN
bracis-33593	232	4	of	of	ADP
bracis-33593	232	5	the	the	DET
bracis-33593	232	6	28th	28th	ADJ
bracis-33593	232	7	international	international	ADJ
bracis-33593	232	8	joint	joint	ADJ
bracis-33593	232	9	conference	conference	NOUN
bracis-33593	232	10	on	on	ADP
bracis-33593	232	11	artificial	artificial	ADJ
bracis-33593	232	12	intelligence	intelligence	NOUN
bracis-33593	232	13	,	,	PUNCT
bracis-33593	232	14	ijcai	ijcai	NOUN
bracis-33593	232	15	2019	2019	NUM
bracis-33593	232	16	,	,	PUNCT
bracis-33593	232	17	pp	pp	ADV
bracis-33593	232	18	.	.	PUNCT
bracis-33593	233	1	3195–3201	3195–3201	NUM
bracis-33593	233	2	.	.	PUNCT
bracis-33593	234	1	aaai	aaai	PROPN
bracis-33593	234	2	press	press	PROPN
bracis-33593	234	3	(	(	PUNCT
bracis-33593	234	4	2019	2019	NUM
bracis-33593	234	5	)	)	PUNCT
bracis-33593	234	6	google	google	NOUN
bracis-33593	234	7	scholar	scholar	NOUN
bracis-33593	234	8	  	  	SPACE
bracis-33593	234	9	rybkin	rybkin	NOUN
bracis-33593	234	10	,	,	PUNCT
bracis-33593	234	11	o.	o.	PROPN
bracis-33593	234	12	,	,	PUNCT
bracis-33593	234	13	zhu	zhu	PROPN
bracis-33593	234	14	,	,	PUNCT
bracis-33593	234	15	c.	c.	PROPN
bracis-33593	234	16	,	,	PUNCT
bracis-33593	234	17	nagabandi	nagabandi	ADV
bracis-33593	234	18	,	,	PUNCT
bracis-33593	234	19	a.	a.	NOUN
bracis-33593	234	20	,	,	PUNCT
bracis-33593	234	21	daniilidis	daniilidi	NOUN
bracis-33593	234	22	,	,	PUNCT
bracis-33593	234	23	k.	k.	PROPN
bracis-33593	234	24	,	,	PUNCT
bracis-33593	234	25	mordatch	mordatch	PROPN
bracis-33593	234	26	,	,	PUNCT
bracis-33593	234	27	i.	i.	PROPN
bracis-33593	234	28	,	,	PUNCT
bracis-33593	234	29	levine	levine	PROPN
bracis-33593	234	30	,	,	PUNCT
bracis-33593	234	31	s.	s.	PROPN
bracis-33593	234	32	:	:	PUNCT
bracis-33593	234	33	model	model	NOUN
bracis-33593	234	34	-	-	PUNCT
bracis-33593	234	35	based	base	VERB
bracis-33593	234	36	reinforcement	reinforcement	NOUN
bracis-33593	234	37	learning	learning	NOUN
bracis-33593	234	38	via	via	ADP
bracis-33593	234	39	latent	latent	NOUN
bracis-33593	234	40	-	-	PUNCT
bracis-33593	234	41	space	space	NOUN
bracis-33593	234	42	collocation	collocation	NOUN
bracis-33593	234	43	,	,	PUNCT
bracis-33593	234	44	vol	vol	NOUN
bracis-33593	234	45	.	.	PUNCT
bracis-33593	234	46	 	 	SPACE
bracis-33593	234	47	139	139	NUM
bracis-33593	234	48	,	,	PUNCT
bracis-33593	234	49	pp	pp	ADJ
bracis-33593	234	50	.	.	PUNCT
bracis-33593	235	1	9190–9201	9190–9201	NUM
bracis-33593	235	2	(	(	PUNCT
bracis-33593	235	3	2021	2021	NUM
bracis-33593	235	4	)	)	PUNCT
bracis-33593	235	5	.	.	PUNCT
bracis-33593	236	1	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85161314975&partnerid=40	https://www.scopus.com/inward/record.uri?eid=2-s2.0-85161314975&partnerid=40	PROPN
bracis-33593	236	2	&	&	CCONJ
bracis-33593	236	3	md5=0f9769abe528c97cbf983f602ecff397	md5=0f9769abe528c97cbf983f602ecff397	PROPN
bracis-33593	236	4	da	da	PROPN
bracis-33593	236	5	 	 	SPACE
bracis-33593	236	6	silva	silva	PROPN
bracis-33593	236	7	,	,	PUNCT
bracis-33593	236	8	b.	b.	PROPN
bracis-33593	236	9	,	,	PUNCT
bracis-33593	236	10	basso	basso	PROPN
bracis-33593	236	11	,	,	PUNCT
bracis-33593	236	12	e.	e.	PROPN
bracis-33593	236	13	,	,	PUNCT
bracis-33593	236	14	bazzan	bazzan	PROPN
bracis-33593	236	15	,	,	PUNCT
bracis-33593	236	16	a.	a.	NOUN
bracis-33593	236	17	,	,	PUNCT
bracis-33593	236	18	engel	engel	PROPN
bracis-33593	236	19	,	,	PUNCT
bracis-33593	236	20	p.	p.	NOUN
bracis-33593	236	21	:	:	PUNCT
bracis-33593	236	22	dealing	deal	VERB
bracis-33593	236	23	with	with	ADP
bracis-33593	236	24	non	non	ADJ
bracis-33593	236	25	-	-	ADJ
bracis-33593	236	26	stationary	stationary	ADJ
bracis-33593	236	27	environments	environment	NOUN
bracis-33593	236	28	using	use	VERB
bracis-33593	236	29	context	context	NOUN
bracis-33593	236	30	detection	detection	NOUN
bracis-33593	236	31	,	,	PUNCT
bracis-33593	236	32	vol	vol	NOUN
bracis-33593	236	33	.	.	PUNCT
bracis-33593	236	34	 	 	SPACE
bracis-33593	236	35	2006	2006	NUM
bracis-33593	236	36	,	,	PUNCT
bracis-33593	236	37	pp	pp	ADV
bracis-33593	236	38	.	.	PUNCT
bracis-33593	237	1	217–224	217–224	NUM
bracis-33593	237	2	(	(	PUNCT
bracis-33593	237	3	2006	2006	NUM
bracis-33593	237	4	)	)	PUNCT
bracis-33593	237	5	.	.	PUNCT
bracis-33593	238	1	https://doi.org/10.1145/1143844.1143872	https://doi.org/10.1145/1143844.1143872	PROPN
bracis-33593	238	2	sutton	sutton	PROPN
bracis-33593	238	3	,	,	PUNCT
bracis-33593	238	4	r.	r.	PROPN
bracis-33593	238	5	,	,	PUNCT
bracis-33593	238	6	barto	barto	PROPN
bracis-33593	238	7	,	,	PUNCT
bracis-33593	238	8	a.	a.	NOUN
bracis-33593	238	9	:	:	PUNCT
bracis-33593	238	10	reinforcement	reinforcement	NOUN
bracis-33593	238	11	learning	learning	NOUN
bracis-33593	238	12	:	:	PUNCT
bracis-33593	238	13	an	an	DET
bracis-33593	238	14	introduction	introduction	NOUN
bracis-33593	238	15	.	.	PUNCT
bracis-33593	239	1	mit	mit	PROPN
bracis-33593	239	2	press	press	PROPN
bracis-33593	239	3	,	,	PUNCT
bracis-33593	239	4	cambridge	cambridge	PROPN
bracis-33593	239	5	(	(	PUNCT
bracis-33593	239	6	2018	2018	NUM
bracis-33593	239	7	)	)	PUNCT
bracis-33593	239	8	math	math	NOUN
bracis-33593	239	9	  	  	SPACE
bracis-33593	239	10	google	google	PROPN
bracis-33593	239	11	scholar	scholar	NOUN
bracis-33593	239	12	  	  	SPACE
bracis-33593	239	13	young	young	ADJ
bracis-33593	239	14	,	,	PUNCT
bracis-33593	239	15	k.	k.	PROPN
bracis-33593	239	16	,	,	PUNCT
bracis-33593	239	17	ramesh	ramesh	PROPN
bracis-33593	239	18	,	,	PUNCT
bracis-33593	239	19	a.	a.	PROPN
bracis-33593	239	20	,	,	PUNCT
bracis-33593	239	21	kirsch	kirsch	PROPN
bracis-33593	239	22	,	,	PUNCT
bracis-33593	239	23	l.	l.	PROPN
bracis-33593	239	24	,	,	PUNCT
bracis-33593	239	25	schmidhuber	schmidhuber	PROPN
bracis-33593	239	26	,	,	PUNCT
bracis-33593	239	27	j.	j.	PROPN
bracis-33593	239	28	:	:	PUNCT
bracis-33593	239	29	the	the	DET
bracis-33593	239	30	benefits	benefit	NOUN
bracis-33593	239	31	of	of	ADP
bracis-33593	239	32	model	model	NOUN
bracis-33593	239	33	-	-	PUNCT
bracis-33593	239	34	based	base	VERB
bracis-33593	239	35	generalization	generalization	NOUN
bracis-33593	239	36	in	in	ADP
bracis-33593	239	37	reinforcement	reinforcement	NOUN
bracis-33593	239	38	learning	learning	NOUN
bracis-33593	239	39	.	.	PUNCT
bracis-33593	240	1	in	in	ADP
bracis-33593	240	2	:	:	PUNCT
bracis-33593	240	3	proceedings	proceeding	NOUN
bracis-33593	240	4	of	of	ADP
bracis-33593	240	5	the	the	DET
bracis-33593	240	6	40th	40th	ADJ
bracis-33593	240	7	international	international	ADJ
bracis-33593	240	8	conference	conference	NOUN
bracis-33593	240	9	on	on	ADP
bracis-33593	240	10	machine	machine	NOUN
bracis-33593	240	11	learning	learning	NOUN
bracis-33593	240	12	.	.	PUNCT
bracis-33593	240	13	icml	icml	PROPN
bracis-33593	240	14	2023	2023	NUM
bracis-33593	240	15	.	.	PUNCT
bracis-33593	241	1	jmlr.org	jmlr.org	X
bracis-33593	241	2	(	(	PUNCT
bracis-33593	241	3	2023	2023	NUM
bracis-33593	241	4	)	)	PUNCT
bracis-33593	241	5	google	google	NOUN
bracis-33593	241	6	scholar	scholar	NOUN
bracis-33593	241	7	  	  	SPACE
bracis-33593	241	8	download	download	NOUN
bracis-33593	241	9	references	reference	NOUN
bracis-33593	241	10	acknowledgments	acknowledgment	NOUN
bracis-33593	241	11	this	this	DET
bracis-33593	241	12	study	study	NOUN
bracis-33593	241	13	was	be	AUX
bracis-33593	241	14	supported	support	VERB
bracis-33593	241	15	in	in	ADP
bracis-33593	241	16	part	part	NOUN
bracis-33593	241	17	by	by	ADP
bracis-33593	241	18	the	the	DET
bracis-33593	241	19	coordenação	coordenação	PROPN
bracis-33593	241	20	de	de	PROPN
bracis-33593	241	21	aperfeiçoamento	aperfeiçoamento	PROPN
bracis-33593	241	22	de	de	X
bracis-33593	241	23	pessoal	pessoal	PROPN
bracis-33593	241	24	de	de	PROPN
bracis-33593	241	25	nível	nível	PROPN
bracis-33593	241	26	superior	superior	ADJ
bracis-33593	241	27	(	(	PUNCT
bracis-33593	241	28	capes	cape	NOUN
bracis-33593	241	29	)	)	PUNCT
bracis-33593	241	30	–	–	PUNCT
bracis-33593	241	31	finance	finance	NOUN
bracis-33593	241	32	code	code	NOUN
bracis-33593	241	33	001	001	NUM
bracis-33593	241	34	,	,	PUNCT
bracis-33593	241	35	and	and	CCONJ
bracis-33593	241	36	by	by	ADP
bracis-33593	241	37	the	the	DET
bracis-33593	241	38	center	center	NOUN
bracis-33593	241	39	for	for	ADP
bracis-33593	241	40	artificial	artificial	ADJ
bracis-33593	241	41	intelligence	intelligence	NOUN
bracis-33593	241	42	(	(	PUNCT
bracis-33593	241	43	c4ai	c4ai	NOUN
bracis-33593	241	44	-	-	PUNCT
bracis-33593	241	45	usp	usp	NOUN
bracis-33593	241	46	)	)	PUNCT
bracis-33593	241	47	,	,	PUNCT
bracis-33593	241	48	with	with	ADP
bracis-33593	241	49	support	support	NOUN
bracis-33593	241	50	by	by	ADP
bracis-33593	241	51	fapesp	fapesp	NOUN
bracis-33593	241	52	(	(	PUNCT
bracis-33593	241	53	grant	grant	VERB
bracis-33593	241	54	#	#	SYM
bracis-33593	241	55	2019/07665	2019/07665	NUM
bracis-33593	241	56	-	-	SYM
bracis-33593	241	57	4	4	NUM
bracis-33593	241	58	)	)	PUNCT
bracis-33593	241	59	and	and	CCONJ
bracis-33593	241	60	by	by	ADP
bracis-33593	241	61	the	the	DET
bracis-33593	241	62	ibm	ibm	PROPN
bracis-33593	241	63	corporation	corporation	NOUN
bracis-33593	241	64	.	.	PUNCT
bracis-33593	242	1	author	author	NOUN
bracis-33593	242	2	information	information	NOUN
bracis-33593	242	3	authors	author	NOUN
bracis-33593	242	4	and	and	CCONJ
bracis-33593	242	5	affiliations	affiliation	NOUN
bracis-33593	242	6	universidade	universidade	PROPN
bracis-33593	242	7	de	de	PROPN
bracis-33593	242	8	são	são	PROPN
bracis-33593	242	9	paulo	paulo	PROPN
bracis-33593	242	10	,	,	PUNCT
bracis-33593	242	11	são	são	PROPN
bracis-33593	242	12	paulo	paulo	PROPN
bracis-33593	242	13	,	,	PUNCT
bracis-33593	242	14	sp	sp	PROPN
bracis-33593	242	15	,	,	PUNCT
bracis-33593	242	16	brazil	brazil	PROPN
bracis-33593	242	17	guilherme	guilherme	PROPN
bracis-33593	242	18	albarrans	albarrans	PROPN
bracis-33593	242	19	 	 	SPACE
bracis-33593	242	20	&	&	CCONJ
bracis-33593	242	21	 	 	SPACE
bracis-33593	242	22	valdinei	valdinei	PROPN
bracis-33593	242	23	freire	freire	PROPN
bracis-33593	242	24	authors	author	NOUN
bracis-33593	242	25	guilherme	guilherme	VERB
bracis-33593	242	26	albarransview	albarransview	NOUN
bracis-33593	242	27	author	author	NOUN
bracis-33593	242	28	publications	publication	NOUN
bracis-33593	242	29	search	search	NOUN
bracis-33593	242	30	author	author	NOUN
bracis-33593	242	31	on	on	ADP
bracis-33593	242	32	:	:	PUNCT
bracis-33593	242	33	pubmed	pubmed	PROPN
bracis-33593	242	34	 	 	SPACE
bracis-33593	242	35	google	google	PROPN
bracis-33593	242	36	scholar	scholar	NOUN
bracis-33593	242	37	valdinei	valdinei	NOUN
bracis-33593	242	38	freireview	freireview	NOUN
bracis-33593	242	39	author	author	NOUN
bracis-33593	242	40	publications	publication	NOUN
bracis-33593	242	41	search	search	NOUN
bracis-33593	242	42	author	author	NOUN
bracis-33593	242	43	on	on	ADP
bracis-33593	242	44	:	:	PUNCT
bracis-33593	242	45	pubmed	pubmed	PROPN
bracis-33593	242	46	 	 	SPACE
bracis-33593	242	47	google	google	PROPN
bracis-33593	242	48	scholar	scholar	NOUN
bracis-33593	242	49	corresponding	correspond	VERB
bracis-33593	242	50	author	author	NOUN
bracis-33593	242	51	correspondence	correspondence	NOUN
bracis-33593	242	52	to	to	ADP
bracis-33593	242	53	guilherme	guilherme	PROPN
bracis-33593	242	54	albarrans	albarrans	PROPN
bracis-33593	242	55	.	.	PUNCT
bracis-33593	243	1	editor	editor	NOUN
bracis-33593	243	2	information	information	NOUN
bracis-33593	243	3	editors	editor	NOUN
bracis-33593	243	4	and	and	CCONJ
bracis-33593	243	5	affiliations	affiliation	NOUN
bracis-33593	243	6	universidade	universidade	VERB
bracis-33593	243	7	federal	federal	PROPN
bracis-33593	243	8	fluminense	fluminense	PROPN
bracis-33593	243	9	,	,	PUNCT
bracis-33593	243	10	niterói	niterói	PROPN
bracis-33593	243	11	,	,	PUNCT
bracis-33593	243	12	brazil	brazil	PROPN
bracis-33593	243	13	aline	aline	PROPN
bracis-33593	243	14	paes	paes	PROPN
bracis-33593	243	15	instituto	instituto	PROPN
bracis-33593	243	16	tecnológico	tecnológico	PROPN
bracis-33593	243	17	de	de	PROPN
bracis-33593	243	18	aeronáutica	aeronáutica	PROPN
bracis-33593	243	19	,	,	PUNCT
bracis-33593	243	20	são	são	PROPN
bracis-33593	243	21	josé	josé	PROPN
bracis-33593	243	22	dos	dos	PROPN
bracis-33593	243	23	campos	campos	PROPN
bracis-33593	243	24	,	,	PUNCT
bracis-33593	243	25	brazil	brazil	PROPN
bracis-33593	243	26	filipe	filipe	PROPN
bracis-33593	243	27	a.	a.	PROPN
bracis-33593	243	28	n.	n.	PROPN
bracis-33593	243	29	verri	verri	PROPN
bracis-33593	243	30	rights	right	NOUN
bracis-33593	243	31	and	and	CCONJ
bracis-33593	243	32	permissions	permission	NOUN
bracis-33593	243	33	reprints	reprint	NOUN
bracis-33593	243	34	and	and	CCONJ
bracis-33593	243	35	permissions	permission	VERB
bracis-33593	243	36	copyright	copyright	NOUN
bracis-33593	243	37	information	information	NOUN
bracis-33593	243	38	©	©	ADP
bracis-33593	243	39	2025	2025	NUM
bracis-33593	243	40	the	the	DET
bracis-33593	243	41	author(s	author(s	NOUN
bracis-33593	243	42	)	)	PUNCT
bracis-33593	243	43	,	,	PUNCT
bracis-33593	243	44	under	under	ADP
bracis-33593	243	45	exclusive	exclusive	ADJ
bracis-33593	243	46	license	license	NOUN
bracis-33593	243	47	to	to	ADP
bracis-33593	243	48	springer	springer	NOUN
bracis-33593	243	49	nature	nature	PROPN
bracis-33593	243	50	switzerland	switzerland	PROPN
bracis-33593	243	51	ag	ag	PROPN
bracis-33593	243	52	about	about	ADP
bracis-33593	243	53	this	this	DET
bracis-33593	243	54	paper	paper	NOUN
bracis-33593	243	55	cite	cite	VERB
bracis-33593	243	56	this	this	DET
bracis-33593	243	57	paper	paper	NOUN
bracis-33593	243	58	albarrans	albarran	NOUN
bracis-33593	243	59	,	,	PUNCT
bracis-33593	243	60	g.	g.	PROPN
bracis-33593	243	61	,	,	PUNCT
bracis-33593	243	62	freire	freire	PROPN
bracis-33593	243	63	,	,	PUNCT
bracis-33593	243	64	v.	v.	PROPN
bracis-33593	243	65	(	(	PUNCT
bracis-33593	243	66	2025	2025	NUM
bracis-33593	243	67	)	)	PUNCT
bracis-33593	243	68	.	.	PUNCT
bracis-33593	244	1	likelihood	likelihood	PROPN
bracis-33593	244	2	estimator	estimator	NOUN
bracis-33593	244	3	for	for	ADP
bracis-33593	244	4	 	 	SPACE
bracis-33593	244	5	multi	multi	PROPN
bracis-33593	244	6	model	model	NOUN
bracis-33593	244	7	-	-	PUNCT
bracis-33593	244	8	based	base	VERB
bracis-33593	244	9	reinforcement	reinforcement	NOUN
bracis-33593	244	10	learning	learning	NOUN
bracis-33593	244	11	.	.	PUNCT
bracis-33593	245	1	in	in	ADP
bracis-33593	245	2	:	:	PUNCT
bracis-33593	245	3	paes	paes	PROPN
bracis-33593	245	4	,	,	PUNCT
bracis-33593	245	5	a.	a.	PROPN
bracis-33593	245	6	,	,	PUNCT
bracis-33593	245	7	verri	verri	NOUN
bracis-33593	245	8	,	,	PUNCT
bracis-33593	245	9	f.a.n	f.a.n	ADJ
bracis-33593	245	10	.	.	PUNCT
bracis-33593	245	11	(	(	PUNCT
bracis-33593	245	12	eds	ed	NOUN
bracis-33593	245	13	)	)	PUNCT
bracis-33593	245	14	intelligent	intelligent	ADJ
bracis-33593	245	15	systems	system	NOUN
bracis-33593	245	16	.	.	PUNCT
bracis-33593	245	17	bracis	bracis	PROPN
bracis-33593	245	18	2024	2024	NUM
bracis-33593	245	19	.	.	PUNCT
bracis-33593	246	1	lecture	lecture	NOUN
bracis-33593	246	2	notes	note	NOUN
bracis-33593	246	3	in	in	ADP
bracis-33593	246	4	computer	computer	NOUN
bracis-33593	246	5	science	science	NOUN
bracis-33593	246	6	(	(	PUNCT
bracis-33593	246	7	)	)	PUNCT
bracis-33593	246	8	,	,	PUNCT
bracis-33593	246	9	vol	vol	NOUN
bracis-33593	246	10	15413	15413	NUM
bracis-33593	246	11	.	.	PUNCT
bracis-33593	247	1	springer	springer	NOUN
bracis-33593	247	2	,	,	PUNCT
bracis-33593	247	3	cham	cham	PROPN
bracis-33593	247	4	.	.	PUNCT
bracis-33593	248	1	https://doi.org/10.1007/978-3-031-79032-4_13	https://doi.org/10.1007/978-3-031-79032-4_13	PRON
bracis-33593	248	2	download	download	NOUN
bracis-33593	248	3	citation	citation	NOUN
bracis-33593	248	4	.ris	.ris	PUNCT
bracis-33593	248	5	.enw	.enw	PROPN
bracis-33593	248	6	.bib	.bib	PUNCT
bracis-33593	249	1	doi	doi	PROPN
bracis-33593	249	2	:	:	PUNCT
bracis-33593	249	3	https://doi.org/10.1007/978-3-031-79032-4_13	https://doi.org/10.1007/978-3-031-79032-4_13	NUM
bracis-33593	249	4	published	publish	VERB
bracis-33593	249	5	:	:	PUNCT
bracis-33593	249	6	30	30	NUM
bracis-33593	249	7	january	january	PROPN
bracis-33593	249	8	2025	2025	NUM
bracis-33593	249	9	publisher	publisher	NOUN
bracis-33593	249	10	name	name	NOUN
bracis-33593	249	11	:	:	PUNCT
bracis-33593	249	12	springer	springer	NOUN
bracis-33593	249	13	,	,	PUNCT
bracis-33593	249	14	cham	cham	PROPN
bracis-33593	249	15	print	print	PROPN
bracis-33593	249	16	isbn	isbn	PROPN
bracis-33593	249	17	:	:	PUNCT
bracis-33593	249	18	978	978	NUM
bracis-33593	249	19	-	-	SYM
bracis-33593	249	20	3	3	NUM
bracis-33593	249	21	-	-	PUNCT
bracis-33593	249	22	031	031	NUM
bracis-33593	249	23	-	-	PUNCT
bracis-33593	249	24	79031	79031	NUM
bracis-33593	249	25	-	-	SYM
bracis-33593	249	26	7	7	NUM
bracis-33593	249	27	online	online	ADJ
bracis-33593	249	28	isbn	isbn	NOUN
bracis-33593	249	29	:	:	PUNCT
bracis-33593	249	30	978	978	NUM
bracis-33593	249	31	-	-	SYM
bracis-33593	249	32	3	3	NUM
bracis-33593	249	33	-	-	PUNCT
bracis-33593	249	34	031	031	NUM
bracis-33593	249	35	-	-	PUNCT
bracis-33593	249	36	79032	79032	NUM
bracis-33593	249	37	-	-	SYM
bracis-33593	249	38	4	4	NUM
bracis-33593	249	39	ebook	ebook	NOUN
bracis-33593	249	40	packages	package	NOUN
bracis-33593	249	41	:	:	PUNCT
bracis-33593	249	42	computer	computer	NOUN
bracis-33593	249	43	sciencecomputer	sciencecomputer	NOUN
bracis-33593	249	44	science	science	NOUN
bracis-33593	249	45	(	(	PUNCT
bracis-33593	249	46	r0	r0	NOUN
bracis-33593	249	47	)	)	PUNCT
bracis-33593	249	48	share	share	VERB
bracis-33593	249	49	this	this	DET
bracis-33593	249	50	paper	paper	NOUN
bracis-33593	249	51	anyone	anyone	PRON
bracis-33593	249	52	you	you	PRON
bracis-33593	249	53	share	share	VERB
bracis-33593	249	54	the	the	DET
bracis-33593	249	55	following	follow	VERB
bracis-33593	249	56	link	link	NOUN
bracis-33593	249	57	with	with	ADP
bracis-33593	249	58	will	will	AUX
bracis-33593	249	59	be	be	AUX
bracis-33593	249	60	able	able	ADJ
bracis-33593	249	61	to	to	PART
bracis-33593	249	62	read	read	VERB
bracis-33593	249	63	this	this	DET
bracis-33593	249	64	content	content	NOUN
bracis-33593	249	65	:	:	PUNCT
bracis-33593	249	66	get	get	VERB
bracis-33593	249	67	shareable	shareable	ADJ
bracis-33593	249	68	linksorry	linksorry	NOUN
bracis-33593	249	69	,	,	PUNCT
bracis-33593	249	70	a	a	DET
bracis-33593	249	71	shareable	shareable	ADJ
bracis-33593	249	72	link	link	NOUN
bracis-33593	249	73	is	be	AUX
bracis-33593	249	74	not	not	PART
bracis-33593	249	75	currently	currently	ADV
bracis-33593	249	76	available	available	ADJ
bracis-33593	249	77	for	for	ADP
bracis-33593	249	78	this	this	DET
bracis-33593	249	79	article	article	NOUN
bracis-33593	249	80	.	.	PUNCT
bracis-33593	250	1	copy	copy	VERB
bracis-33593	250	2	shareable	shareable	ADJ
bracis-33593	250	3	link	link	NOUN
bracis-33593	250	4	to	to	PART
bracis-33593	250	5	clipboard	clipboard	NOUN
bracis-33593	250	6	provided	provide	VERB
bracis-33593	250	7	by	by	ADP
bracis-33593	250	8	the	the	DET
bracis-33593	250	9	springer	springer	NOUN
bracis-33593	250	10	nature	nature	PROPN
bracis-33593	250	11	sharedit	sharedit	PROPN
bracis-33593	250	12	content	content	NOUN
bracis-33593	250	13	-	-	PUNCT
bracis-33593	250	14	sharing	share	VERB
bracis-33593	250	15	initiative	initiative	NOUN
bracis-33593	250	16	keywords	keyword	NOUN
bracis-33593	250	17	model	model	VERB
bracis-33593	250	18	based	base	VERB
bracis-33593	250	19	reinforcement	reinforcement	NOUN
bracis-33593	250	20	learning	learning	NOUN
bracis-33593	250	21	planning	planning	NOUN
bracis-33593	250	22	likelihood	likelihood	NOUN
bracis-33593	250	23	estimator	estimator	NOUN
bracis-33593	250	24	publish	publish	VERB
bracis-33593	250	25	with	with	ADP
bracis-33593	250	26	us	us	PROPN
bracis-33593	250	27	policies	policy	NOUN
bracis-33593	250	28	and	and	CCONJ
bracis-33593	250	29	ethics	ethic	NOUN
bracis-33593	250	30	search	search	NOUN
bracis-33593	250	31	search	search	NOUN
bracis-33593	250	32	by	by	ADP
bracis-33593	250	33	keyword	keyword	NOUN
bracis-33593	250	34	or	or	CCONJ
bracis-33593	250	35	author	author	NOUN
bracis-33593	250	36	search	search	NOUN
bracis-33593	250	37	navigation	navigation	NOUN
bracis-33593	250	38	find	find	VERB
bracis-33593	250	39	a	a	DET
bracis-33593	250	40	journal	journal	NOUN
bracis-33593	250	41	publish	publish	VERB
bracis-33593	250	42	with	with	ADP
bracis-33593	250	43	us	we	PRON
bracis-33593	250	44	track	track	VERB
bracis-33593	250	45	your	your	PRON
bracis-33593	250	46	research	research	NOUN
bracis-33593	250	47	discover	discover	VERB
bracis-33593	250	48	content	content	NOUN
bracis-33593	250	49	journals	journal	NOUN
bracis-33593	250	50	a	a	DET
bracis-33593	250	51	-	-	PUNCT
bracis-33593	250	52	z	z	NOUN
bracis-33593	250	53	books	book	NOUN
bracis-33593	250	54	a	a	DET
bracis-33593	250	55	-	-	PUNCT
bracis-33593	250	56	z	z	NOUN
bracis-33593	250	57	publish	publish	NOUN
bracis-33593	250	58	with	with	ADP
bracis-33593	250	59	us	us	PROPN
bracis-33593	250	60	journal	journal	PROPN
bracis-33593	250	61	finder	finder	PROPN
bracis-33593	250	62	publish	publish	VERB
bracis-33593	250	63	your	your	PRON
bracis-33593	250	64	research	research	NOUN
bracis-33593	250	65	language	language	NOUN
bracis-33593	250	66	editing	edit	VERB
bracis-33593	250	67	open	open	ADJ
bracis-33593	250	68	access	access	NOUN
bracis-33593	250	69	publishing	publishing	NOUN
bracis-33593	250	70	products	product	NOUN
bracis-33593	250	71	and	and	CCONJ
bracis-33593	250	72	services	service	NOUN
bracis-33593	250	73	our	our	PRON
bracis-33593	250	74	products	product	NOUN
bracis-33593	250	75	librarians	librarian	VERB
bracis-33593	250	76	societies	society	NOUN
bracis-33593	250	77	partners	partner	NOUN
bracis-33593	250	78	and	and	CCONJ
bracis-33593	250	79	advertisers	advertiser	NOUN
bracis-33593	250	80	our	our	PRON
bracis-33593	250	81	brands	brand	NOUN
bracis-33593	250	82	springer	springer	NOUN
bracis-33593	250	83	nature	nature	PROPN
bracis-33593	250	84	portfolio	portfolio	PROPN
bracis-33593	250	85	bmc	bmc	PROPN
bracis-33593	250	86	palgrave	palgrave	PROPN
bracis-33593	250	87	macmillan	macmillan	PROPN
bracis-33593	250	88	apress	apress	PROPN
bracis-33593	250	89	discover	discover	VERB
bracis-33593	250	90	your	your	PRON
bracis-33593	250	91	privacy	privacy	NOUN
bracis-33593	250	92	choices	choice	NOUN
bracis-33593	250	93	/	/	SYM
bracis-33593	250	94	manage	manage	NOUN
bracis-33593	250	95	cookies	cookie	NOUN
bracis-33593	250	96	your	your	PRON
bracis-33593	250	97	us	us	PROPN
bracis-33593	251	1	state	state	NOUN
bracis-33593	251	2	privacy	privacy	NOUN
bracis-33593	251	3	rights	right	NOUN
bracis-33593	251	4	accessibility	accessibility	NOUN
bracis-33593	251	5	statement	statement	NOUN
bracis-33593	251	6	terms	term	NOUN
bracis-33593	251	7	and	and	CCONJ
bracis-33593	251	8	conditions	condition	NOUN
bracis-33593	251	9	privacy	privacy	NOUN
bracis-33593	251	10	policy	policy	NOUN
bracis-33593	251	11	help	help	NOUN
bracis-33593	251	12	and	and	CCONJ
bracis-33593	251	13	support	support	VERB
bracis-33593	251	14	legal	legal	ADJ
bracis-33593	251	15	notice	notice	NOUN
bracis-33593	251	16	cancel	cancel	VERB
bracis-33593	251	17	contracts	contract	NOUN
bracis-33593	251	18	here	here	ADV
bracis-33593	251	19	129.74.145.123	129.74.145.123	NUM
bracis-33593	251	20	hesburgh	hesburgh	PROPN
bracis-33593	251	21	library	library	PROPN
bracis-33593	251	22	er	er	INTJ
bracis-33593	251	23	unit	unit	NOUN
bracis-33593	251	24	(	(	PUNCT
bracis-33593	251	25	3005732405	3005732405	NUM
bracis-33593	251	26	)	)	PUNCT
bracis-33593	251	27	northeast	northeast	ADJ
bracis-33593	251	28	research	research	NOUN
bracis-33593	251	29	libraries	library	NOUN
bracis-33593	251	30	(	(	PUNCT
bracis-33593	251	31	nerl	nerl	PROPN
bracis-33593	251	32	)	)	PUNCT
bracis-33593	251	33	(	(	PUNCT
bracis-33593	251	34	8200828607	8200828607	NUM
bracis-33593	251	35	)	)	PUNCT
bracis-33593	251	36	nerl	nerl	VERB
bracis-33593	251	37	ta	ta	X
bracis-33593	251	38	account	account	NOUN
bracis-33593	251	39	(	(	PUNCT
bracis-33593	251	40	3006206169	3006206169	NUM
bracis-33593	251	41	)	)	PUNCT
bracis-33593	251	42	university	university	NOUN
bracis-33593	251	43	of	of	ADP
bracis-33593	251	44	notre	notre	PROPN
bracis-33593	251	45	dame	dame	PROPN
bracis-33593	251	46	hesburgh	hesburgh	PROPN
bracis-33593	251	47	library	library	NOUN
bracis-33593	251	48	(	(	PUNCT
bracis-33593	251	49	3000184373	3000184373	NUM
bracis-33593	251	50	)	)	PUNCT
bracis-33593	252	1	©	©	ADP
bracis-33593	252	2	2025	2025	NUM
bracis-33593	252	3	springer	springer	NOUN
bracis-33593	252	4	nature	nature	NOUN