id	sid	tid	token	lemma	pos
cana-6094	1	1	communications	communication	NOUN
cana-6094	1	2	on	on	ADP
cana-6094	1	3	applied	apply	VERB
cana-6094	1	4	nonlinear	nonlinear	ADJ
cana-6094	1	5	analysis	analysis	NOUN
cana-6094	1	6	issn	issn	NOUN
cana-6094	1	7	:	:	PUNCT
cana-6094	1	8	1074	1074	NUM
cana-6094	1	9	-	-	PUNCT
cana-6094	1	10	133x	133x	NUM
cana-6094	1	11	vol	vol	NOUN
cana-6094	1	12	29	29	NUM
cana-6094	1	13	no	no	NOUN
cana-6094	1	14	.	.	NOUN
cana-6094	1	15	4	4	NUM
cana-6094	1	16	(	(	PUNCT
cana-6094	1	17	2022	2022	NUM
cana-6094	1	18	)	)	PUNCT
cana-6094	1	19	27	27	NUM
cana-6094	1	20	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	1	21	graph	graph	NOUN
cana-6094	1	22	neural	neural	ADJ
cana-6094	1	23	networks	network	NOUN
cana-6094	1	24	for	for	ADP
cana-6094	1	25	the	the	DET
cana-6094	1	26	analysis	analysis	NOUN
cana-6094	1	27	of	of	ADP
cana-6094	1	28	complex	complex	ADJ
cana-6094	1	29	mathematical	mathematical	ADJ
cana-6094	1	30	structures	structure	NOUN
cana-6094	1	31	machhindranath	machhindranath	PROPN
cana-6094	1	32	m.	m.	PROPN
cana-6094	1	33	dhane1	dhane1	PROPN
cana-6094	1	34	*	*	PROPN
cana-6094	1	35	,	,	PUNCT
cana-6094	1	36	shobhankumar	shobhankumar	VERB
cana-6094	1	37	d.m2	d.m2	NOUN
cana-6094	1	38	,	,	PUNCT
cana-6094	1	39	vanaja	vanaja	PROPN
cana-6094	1	40	v3	v3	PROPN
cana-6094	1	41	,	,	PUNCT
cana-6094	1	42	shailaja	shailaja	PROPN
cana-6094	1	43	m4	m4	PROPN
cana-6094	1	44	1*associate	1*associate	PROPN
cana-6094	1	45	professor	professor	NOUN
cana-6094	1	46	,	,	PUNCT
cana-6094	1	47	department	department	NOUN
cana-6094	1	48	of	of	ADP
cana-6094	1	49	mathematics	mathematics	PROPN
cana-6094	1	50	,	,	PUNCT
cana-6094	1	51	gfgc	gfgc	PROPN
cana-6094	1	52	,	,	PUNCT
cana-6094	1	53	yelahanka	yelahanka	PROPN
cana-6094	1	54	,	,	PUNCT
cana-6094	1	55	bengaluru	bengaluru	PROPN
cana-6094	1	56	,	,	PUNCT
cana-6094	1	57	karnataka	karnataka	PROPN
cana-6094	1	58	,	,	PUNCT
cana-6094	1	59	india	india	PROPN
cana-6094	1	60	;	;	PUNCT
cana-6094	1	61	email	email	NOUN
cana-6094	1	62	:	:	PUNCT
cana-6094	1	63	drdhani.99dce@gmail.com	drdhani.99dce@gmail.com	X
cana-6094	2	1	2associate	2associate	NUM
cana-6094	2	2	professor	professor	NOUN
cana-6094	2	3	,	,	PUNCT
cana-6094	2	4	department	department	NOUN
cana-6094	2	5	of	of	ADP
cana-6094	2	6	mathematics	mathematics	PROPN
cana-6094	2	7	,	,	PUNCT
cana-6094	2	8	maharani	maharani	PROPN
cana-6094	2	9	’s	’s	PROPN
cana-6094	2	10	science	science	PROPN
cana-6094	2	11	college	college	PROPN
cana-6094	2	12	for	for	ADP
cana-6094	2	13	women	woman	NOUN
cana-6094	2	14	,	,	PUNCT
cana-6094	2	15	palace	palace	NOUN
cana-6094	2	16	road	road	NOUN
cana-6094	2	17	,	,	PUNCT
cana-6094	2	18	bengaluru	bengaluru	PROPN
cana-6094	2	19	,	,	PUNCT
cana-6094	2	20	karnataka	karnataka	PROPN
cana-6094	2	21	,	,	PUNCT
cana-6094	2	22	india	india	PROPN
cana-6094	2	23	;	;	PUNCT
cana-6094	2	24	email:shobhankumardm@gmail.com	email:shobhankumardm@gmail.com	X
cana-6094	2	25	3	3	NUM
cana-6094	2	26	associate	associate	NOUN
cana-6094	2	27	professor	professor	NOUN
cana-6094	2	28	,	,	PUNCT
cana-6094	2	29	department	department	NOUN
cana-6094	2	30	of	of	ADP
cana-6094	2	31	mathematics	mathematics	PROPN
cana-6094	2	32	,	,	PUNCT
cana-6094	2	33	gfgc	gfgc	PROPN
cana-6094	2	34	,	,	PUNCT
cana-6094	2	35	yelahanka	yelahanka	PROPN
cana-6094	2	36	,	,	PUNCT
cana-6094	2	37	bengaluru	bengaluru	PROPN
cana-6094	2	38	,	,	PUNCT
cana-6094	2	39	karnataka	karnataka	PROPN
cana-6094	2	40	,	,	PUNCT
cana-6094	2	41	india	india	PROPN
cana-6094	2	42	;	;	PUNCT
cana-6094	2	43	email	email	NOUN
cana-6094	2	44	:	:	PUNCT
cana-6094	2	45	vanajaksr94@yahoo.com	vanajaksr94@yahoo.com	PROPN
cana-6094	2	46	4	4	NUM
cana-6094	2	47	associate	associate	NOUN
cana-6094	2	48	professor	professor	NOUN
cana-6094	2	49	,	,	PUNCT
cana-6094	2	50	department	department	NOUN
cana-6094	2	51	of	of	ADP
cana-6094	2	52	mathematics	mathematics	PROPN
cana-6094	2	53	,	,	PUNCT
cana-6094	2	54	gfgc	gfgc	NOUN
cana-6094	2	55	,	,	PUNCT
cana-6094	2	56	vijayanagar	vijayanagar	NOUN
cana-6094	2	57	,	,	PUNCT
cana-6094	2	58	bengaluru	bengaluru	PROPN
cana-6094	2	59	,	,	PUNCT
cana-6094	2	60	karnataka	karnataka	PROPN
cana-6094	2	61	,	,	PUNCT
cana-6094	2	62	india	india	PROPN
cana-6094	2	63	;	;	PUNCT
cana-6094	2	64	email	email	NOUN
cana-6094	2	65	:	:	PUNCT
cana-6094	2	66	shaila.ac55@yahoo.com	shaila.ac55@yahoo.com	X
cana-6094	2	67	*	*	PUNCT
cana-6094	2	68	corresponding	correspond	VERB
cana-6094	2	69	author	author	NOUN
cana-6094	2	70	:	:	PUNCT
cana-6094	2	71	drdhani.99dce@gmail.com	drdhani.99dce@gmail.com	X
cana-6094	2	72	email	email	NOUN
cana-6094	2	73	of	of	ADP
cana-6094	2	74	co-author:shobhankumardm@gmail.com	co-author:shobhankumardm@gmail.com	NOUN
cana-6094	2	75	article	article	NOUN
cana-6094	2	76	history	history	NOUN
cana-6094	2	77	:	:	PUNCT
cana-6094	2	78	received	receive	VERB
cana-6094	2	79	:	:	PUNCT
cana-6094	2	80	09	09	NUM
cana-6094	2	81	-	-	SYM
cana-6094	2	82	06	06	NUM
cana-6094	2	83	-	-	PUNCT
cana-6094	2	84	2022	2022	NUM
cana-6094	2	85	revised	revise	VERB
cana-6094	2	86	:	:	PUNCT
cana-6094	2	87	22	22	NUM
cana-6094	2	88	-	-	SYM
cana-6094	2	89	07	07	NUM
cana-6094	2	90	-	-	PUNCT
cana-6094	2	91	2022	2022	NUM
cana-6094	2	92	accepted	accept	VERB
cana-6094	2	93	:	:	PUNCT
cana-6094	2	94	29	29	NUM
cana-6094	2	95	-	-	SYM
cana-6094	2	96	08	08	NUM
cana-6094	2	97	-	-	PUNCT
cana-6094	2	98	2022	2022	NUM
cana-6094	2	99	abstract	abstract	NOUN
cana-6094	2	100	:	:	PUNCT
cana-6094	2	101	graph	graph	NOUN
cana-6094	2	102	neural	neural	ADJ
cana-6094	2	103	networks	network	NOUN
cana-6094	2	104	(	(	PUNCT
cana-6094	2	105	gnns	gnns	NOUN
cana-6094	2	106	)	)	PUNCT
cana-6094	2	107	have	have	AUX
cana-6094	2	108	emerged	emerge	VERB
cana-6094	2	109	as	as	ADP
cana-6094	2	110	a	a	DET
cana-6094	2	111	powerful	powerful	ADJ
cana-6094	2	112	framework	framework	NOUN
cana-6094	2	113	for	for	ADP
cana-6094	2	114	analyzing	analyze	VERB
cana-6094	2	115	complex	complex	ADJ
cana-6094	2	116	mathematical	mathematical	ADJ
cana-6094	2	117	structures	structure	NOUN
cana-6094	2	118	that	that	PRON
cana-6094	2	119	can	can	AUX
cana-6094	2	120	be	be	AUX
cana-6094	2	121	represented	represent	VERB
cana-6094	2	122	as	as	ADP
cana-6094	2	123	graphs	graph	NOUN
cana-6094	2	124	,	,	PUNCT
cana-6094	2	125	including	include	VERB
cana-6094	2	126	algebraic	algebraic	ADJ
cana-6094	2	127	objects	object	NOUN
cana-6094	2	128	,	,	PUNCT
cana-6094	2	129	geometric	geometric	ADJ
cana-6094	2	130	spaces	space	NOUN
cana-6094	2	131	,	,	PUNCT
cana-6094	2	132	and	and	CCONJ
cana-6094	2	133	combinatorial	combinatorial	ADJ
cana-6094	2	134	structures	structure	NOUN
cana-6094	2	135	.	.	PUNCT
cana-6094	3	1	this	this	DET
cana-6094	3	2	comprehensive	comprehensive	ADJ
cana-6094	3	3	review	review	NOUN
cana-6094	3	4	examines	examine	VERB
cana-6094	3	5	the	the	DET
cana-6094	3	6	theoretical	theoretical	ADJ
cana-6094	3	7	foundations	foundation	NOUN
cana-6094	3	8	,	,	PUNCT
cana-6094	3	9	architectural	architectural	ADJ
cana-6094	3	10	variants	variant	NOUN
cana-6094	3	11	,	,	PUNCT
cana-6094	3	12	and	and	CCONJ
cana-6094	3	13	applications	application	NOUN
cana-6094	3	14	of	of	ADP
cana-6094	3	15	gnns	gnns	NOUN
cana-6094	3	16	to	to	PART
cana-6094	3	17	mathematical	mathematical	ADJ
cana-6094	3	18	analysis	analysis	NOUN
cana-6094	3	19	,	,	PUNCT
cana-6094	3	20	highlighting	highlight	VERB
cana-6094	3	21	how	how	SCONJ
cana-6094	3	22	these	these	DET
cana-6094	3	23	networks	network	NOUN
cana-6094	3	24	effectively	effectively	ADV
cana-6094	3	25	capture	capture	VERB
cana-6094	3	26	the	the	DET
cana-6094	3	27	relational	relational	ADJ
cana-6094	3	28	inductive	inductive	ADJ
cana-6094	3	29	biases	bias	NOUN
cana-6094	3	30	inherent	inherent	ADJ
cana-6094	3	31	in	in	ADP
cana-6094	3	32	mathematical	mathematical	ADJ
cana-6094	3	33	systems	system	NOUN
cana-6094	3	34	.	.	PUNCT
cana-6094	4	1	we	we	PRON
cana-6094	4	2	explore	explore	VERB
cana-6094	4	3	how	how	SCONJ
cana-6094	4	4	gnns	gnns	ADJ
cana-6094	4	5	generalize	generalize	VERB
cana-6094	4	6	traditional	traditional	ADJ
cana-6094	4	7	neural	neural	ADJ
cana-6094	4	8	network	network	NOUN
cana-6094	4	9	operations	operation	NOUN
cana-6094	4	10	to	to	ADP
cana-6094	4	11	graph	graph	NOUN
cana-6094	4	12	-	-	PUNCT
cana-6094	4	13	structured	structure	VERB
cana-6094	4	14	data	datum	NOUN
cana-6094	4	15	through	through	ADP
cana-6094	4	16	message	message	NOUN
cana-6094	4	17	passing	passing	NOUN
cana-6094	4	18	,	,	PUNCT
cana-6094	4	19	aggregation	aggregation	NOUN
cana-6094	4	20	,	,	PUNCT
cana-6094	4	21	and	and	CCONJ
cana-6094	4	22	update	update	NOUN
cana-6094	4	23	mechanisms	mechanism	NOUN
cana-6094	4	24	,	,	PUNCT
cana-6094	4	25	enabling	enable	VERB
cana-6094	4	26	them	they	PRON
cana-6094	4	27	to	to	PART
cana-6094	4	28	process	process	VERB
cana-6094	4	29	mathematical	mathematical	ADJ
cana-6094	4	30	structures	structure	NOUN
cana-6094	4	31	with	with	ADP
cana-6094	4	32	irregular	irregular	ADJ
cana-6094	4	33	and	and	CCONJ
cana-6094	4	34	non	non	ADJ
cana-6094	4	35	-	-	ADJ
cana-6094	4	36	euclidean	euclidean	ADJ
cana-6094	4	37	properties	property	NOUN
cana-6094	4	38	.	.	PUNCT
cana-6094	5	1	the	the	DET
cana-6094	5	2	paper	paper	NOUN
cana-6094	5	3	surveys	survey	NOUN
cana-6094	5	4	specialized	specialized	PROPN
cana-6094	5	5	gnn	gnn	PROPN
cana-6094	5	6	architectures	architecture	NOUN
cana-6094	5	7	including	include	VERB
cana-6094	5	8	spectral	spectral	ADJ
cana-6094	5	9	convolutional	convolutional	ADJ
cana-6094	5	10	networks	network	NOUN
cana-6094	5	11	,	,	PUNCT
cana-6094	5	12	attention	attention	NOUN
cana-6094	5	13	-	-	PUNCT
cana-6094	5	14	based	base	VERB
cana-6094	5	15	models	model	NOUN
cana-6094	5	16	,	,	PUNCT
cana-6094	5	17	and	and	CCONJ
cana-6094	5	18	geometrically	geometrically	ADV
cana-6094	5	19	equivariant	equivariant	ADJ
cana-6094	5	20	networks	network	NOUN
cana-6094	5	21	that	that	PRON
cana-6094	5	22	preserve	preserve	VERB
cana-6094	5	23	mathematical	mathematical	ADJ
cana-6094	5	24	symmetries	symmetry	NOUN
cana-6094	5	25	.	.	PUNCT
cana-6094	6	1	through	through	ADP
cana-6094	6	2	detailed	detailed	ADJ
cana-6094	6	3	case	case	NOUN
cana-6094	6	4	studies	study	NOUN
cana-6094	6	5	in	in	ADP
cana-6094	6	6	group	group	NOUN
cana-6094	6	7	theory	theory	NOUN
cana-6094	6	8	,	,	PUNCT
cana-6094	6	9	topology	topology	NOUN
cana-6094	6	10	,	,	PUNCT
cana-6094	6	11	combinatorics	combinatoric	NOUN
cana-6094	6	12	,	,	PUNCT
cana-6094	6	13	and	and	CCONJ
cana-6094	6	14	mathematical	mathematical	ADJ
cana-6094	6	15	physics	physics	NOUN
cana-6094	6	16	,	,	PUNCT
cana-6094	6	17	we	we	PRON
cana-6094	6	18	demonstrate	demonstrate	VERB
cana-6094	6	19	how	how	SCONJ
cana-6094	6	20	gnns	gnns	NOUN
cana-6094	6	21	can	can	AUX
cana-6094	6	22	discover	discover	VERB
cana-6094	6	23	patterns	pattern	NOUN
cana-6094	6	24	,	,	PUNCT
cana-6094	6	25	predict	predict	VERB
cana-6094	6	26	properties	property	NOUN
cana-6094	6	27	,	,	PUNCT
cana-6094	6	28	and	and	CCONJ
cana-6094	6	29	generate	generate	VERB
cana-6094	6	30	conjectures	conjecture	VERB
cana-6094	6	31	about	about	ADP
cana-6094	6	32	mathematical	mathematical	ADJ
cana-6094	6	33	structures	structure	NOUN
cana-6094	6	34	.	.	PUNCT
cana-6094	7	1	the	the	DET
cana-6094	7	2	review	review	NOUN
cana-6094	7	3	also	also	ADV
cana-6094	7	4	addresses	address	VERB
cana-6094	7	5	current	current	ADJ
cana-6094	7	6	challenges	challenge	NOUN
cana-6094	7	7	and	and	CCONJ
cana-6094	7	8	limitations	limitation	NOUN
cana-6094	7	9	while	while	SCONJ
cana-6094	7	10	outlining	outline	VERB
cana-6094	7	11	promising	promise	VERB
cana-6094	7	12	research	research	NOUN
cana-6094	7	13	directions	direction	NOUN
cana-6094	7	14	for	for	ADP
cana-6094	7	15	advancing	advance	VERB
cana-6094	7	16	the	the	DET
cana-6094	7	17	integration	integration	NOUN
cana-6094	7	18	of	of	ADP
cana-6094	7	19	graph	graph	NOUN
cana-6094	7	20	representation	representation	NOUN
cana-6094	7	21	learning	learn	VERB
cana-6094	7	22	with	with	ADP
cana-6094	7	23	mathematical	mathematical	ADJ
cana-6094	7	24	reasoning	reasoning	NOUN
cana-6094	7	25	.	.	PUNCT
cana-6094	8	1	keywors	keywor	NOUN
cana-6094	8	2	:	:	PUNCT
cana-6094	8	3	graph	graph	VERB
cana-6094	8	4	neural	neural	ADJ
cana-6094	8	5	networks	network	NOUN
cana-6094	8	6	,	,	PUNCT
cana-6094	8	7	geometric	geometric	ADJ
cana-6094	8	8	deep	deep	ADJ
cana-6094	8	9	learning	learning	NOUN
cana-6094	8	10	,	,	PUNCT
cana-6094	8	11	message	message	NOUN
cana-6094	8	12	passing	pass	VERB
cana-6094	8	13	neural	neural	ADJ
cana-6094	8	14	networks	network	NOUN
cana-6094	8	15	,	,	PUNCT
cana-6094	8	16	relational	relational	ADJ
cana-6094	8	17	inductive	inductive	ADJ
cana-6094	8	18	biases	bias	NOUN
cana-6094	8	19	.	.	PUNCT
cana-6094	9	1	1.introduction	1.introduction	NUM
cana-6094	9	2	complex	complex	ADJ
cana-6094	9	3	mathematical	mathematical	ADJ
cana-6094	9	4	structures	structure	NOUN
cana-6094	9	5	including	include	VERB
cana-6094	9	6	groups	group	NOUN
cana-6094	9	7	,	,	PUNCT
cana-6094	9	8	rings	ring	NOUN
cana-6094	9	9	,	,	PUNCT
cana-6094	9	10	topological	topological	ADJ
cana-6094	9	11	spaces	space	NOUN
cana-6094	9	12	,	,	PUNCT
cana-6094	9	13	and	and	CCONJ
cana-6094	9	14	combinatorial	combinatorial	ADJ
cana-6094	9	15	objects	object	NOUN
cana-6094	9	16	form	form	VERB
cana-6094	9	17	the	the	DET
cana-6094	9	18	foundation	foundation	NOUN
cana-6094	9	19	of	of	ADP
cana-6094	9	20	modern	modern	ADJ
cana-6094	9	21	mathematics	mathematic	NOUN
cana-6094	9	22	and	and	CCONJ
cana-6094	9	23	its	its	PRON
cana-6094	9	24	applications	application	NOUN
cana-6094	9	25	across	across	ADP
cana-6094	9	26	scientific	scientific	ADJ
cana-6094	9	27	disciplines	discipline	NOUN
cana-6094	9	28	.	.	PUNCT
cana-6094	10	1	traditional	traditional	ADJ
cana-6094	10	2	approaches	approach	NOUN
cana-6094	10	3	to	to	ADP
cana-6094	10	4	analyzing	analyze	VERB
cana-6094	10	5	these	these	DET
cana-6094	10	6	structures	structure	NOUN
cana-6094	10	7	have	have	AUX
cana-6094	10	8	relied	rely	VERB
cana-6094	10	9	on	on	ADP
cana-6094	10	10	symbolic	symbolic	ADJ
cana-6094	10	11	computation	computation	NOUN
cana-6094	10	12	,	,	PUNCT
cana-6094	10	13	theorem	theorem	ADJ
cana-6094	10	14	proving	proving	NOUN
cana-6094	10	15	,	,	PUNCT
cana-6094	10	16	and	and	CCONJ
cana-6094	10	17	human	human	ADJ
cana-6094	10	18	intuition	intuition	NOUN
cana-6094	10	19	,	,	PUNCT
cana-6094	10	20	which	which	PRON
cana-6094	10	21	often	often	ADV
cana-6094	10	22	struggle	struggle	VERB
cana-6094	10	23	with	with	ADP
cana-6094	10	24	the	the	DET
cana-6094	10	25	complexity	complexity	NOUN
cana-6094	10	26	and	and	CCONJ
cana-6094	10	27	scale	scale	NOUN
cana-6094	10	28	of	of	ADP
cana-6094	10	29	many	many	ADJ
cana-6094	10	30	contemporary	contemporary	ADJ
cana-6094	10	31	mathematical	mathematical	ADJ
cana-6094	10	32	problems	problem	NOUN
cana-6094	10	33	.	.	PUNCT
cana-6094	11	1	the	the	DET
cana-6094	11	2	emergence	emergence	NOUN
cana-6094	11	3	of	of	ADP
cana-6094	11	4	machine	machine	NOUN
cana-6094	11	5	learning	learning	NOUN
cana-6094	11	6	methods	method	NOUN
cana-6094	11	7	,	,	PUNCT
cana-6094	11	8	particularly	particularly	ADV
cana-6094	11	9	deep	deep	ADJ
cana-6094	11	10	learning	learning	NOUN
cana-6094	11	11	,	,	PUNCT
cana-6094	11	12	has	have	AUX
cana-6094	11	13	introduced	introduce	VERB
cana-6094	11	14	new	new	ADJ
cana-6094	11	15	paradigms	paradigm	NOUN
cana-6094	11	16	for	for	ADP
cana-6094	11	17	mathematical	mathematical	ADJ
cana-6094	11	18	discovery	discovery	NOUN
cana-6094	11	19	by	by	ADP
cana-6094	11	20	leveraging	leverage	VERB
cana-6094	11	21	pattern	pattern	NOUN
cana-6094	11	22	recognition	recognition	NOUN
cana-6094	11	23	and	and	CCONJ
cana-6094	11	24	statistical	statistical	ADJ
cana-6094	11	25	learning	learning	NOUN
cana-6094	11	26	to	to	PART
cana-6094	11	27	complement	complement	VERB
cana-6094	11	28	formal	formal	ADJ
cana-6094	11	29	reasoning	reasoning	NOUN
cana-6094	11	30	.	.	PUNCT
cana-6094	12	1	graph	graph	NOUN
cana-6094	12	2	neural	neural	ADJ
cana-6094	12	3	mailto:drdhani.99dce@gmail.com	mailto:drdhani.99dce@gmail.com	X
cana-6094	12	4	mailto:shobhankumardm@gmail.com	mailto:shobhankumardm@gmail.com	X
cana-6094	12	5	mailto:vanajaksr94@yahoo.com	mailto:vanajaksr94@yahoo.com	PROPN
cana-6094	12	6	mailto:shaila.ac55@yahoo.com	mailto:shaila.ac55@yahoo.com	PUNCT
cana-6094	13	1	mailto:drdhani.99dce@gmail.com	mailto:drdhani.99dce@gmail.com	PROPN
cana-6094	13	2	mailto:shobhankumardm@gmail.com	mailto:shobhankumardm@gmail.com	X
cana-6094	13	3	communications	communication	NOUN
cana-6094	13	4	on	on	ADP
cana-6094	13	5	applied	apply	VERB
cana-6094	13	6	nonlinear	nonlinear	ADJ
cana-6094	13	7	analysis	analysis	NOUN
cana-6094	13	8	issn	issn	NOUN
cana-6094	13	9	:	:	PUNCT
cana-6094	13	10	1074	1074	NUM
cana-6094	13	11	-	-	PUNCT
cana-6094	13	12	133x	133x	NUM
cana-6094	13	13	vol	vol	NOUN
cana-6094	13	14	29	29	NUM
cana-6094	13	15	no	no	NOUN
cana-6094	13	16	.	.	NOUN
cana-6094	13	17	4	4	NUM
cana-6094	13	18	(	(	PUNCT
cana-6094	13	19	2022	2022	NUM
cana-6094	13	20	)	)	PUNCT
cana-6094	13	21	28	28	NUM
cana-6094	14	1	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	14	2	networks	network	NOUN
cana-6094	14	3	(	(	PUNCT
cana-6094	14	4	gnns	gnns	NOUN
cana-6094	14	5	)	)	PUNCT
cana-6094	14	6	represent	represent	VERB
cana-6094	14	7	a	a	DET
cana-6094	14	8	particularly	particularly	ADV
cana-6094	14	9	promising	promising	ADJ
cana-6094	14	10	approach	approach	NOUN
cana-6094	14	11	for	for	ADP
cana-6094	14	12	mathematical	mathematical	ADJ
cana-6094	14	13	analysis	analysis	NOUN
cana-6094	14	14	due	due	ADP
cana-6094	14	15	to	to	ADP
cana-6094	14	16	their	their	PRON
cana-6094	14	17	ability	ability	NOUN
cana-6094	14	18	to	to	PART
cana-6094	14	19	process	process	VERB
cana-6094	14	20	relational	relational	ADJ
cana-6094	14	21	data	datum	NOUN
cana-6094	14	22	and	and	CCONJ
cana-6094	14	23	capture	capture	VERB
cana-6094	14	24	structural	structural	ADJ
cana-6094	14	25	patterns	pattern	NOUN
cana-6094	14	26	.	.	PUNCT
cana-6094	15	1	mathematical	mathematical	ADJ
cana-6094	15	2	structures	structure	NOUN
cana-6094	15	3	naturally	naturally	ADV
cana-6094	15	4	exhibit	exhibit	VERB
cana-6094	15	5	graph	graph	NOUN
cana-6094	15	6	-	-	PUNCT
cana-6094	15	7	like	like	ADJ
cana-6094	15	8	properties	property	NOUN
cana-6094	15	9	:	:	PUNCT
cana-6094	15	10	algebraic	algebraic	ADJ
cana-6094	15	11	structures	structure	NOUN
cana-6094	15	12	can	can	AUX
cana-6094	15	13	be	be	AUX
cana-6094	15	14	represented	represent	VERB
cana-6094	15	15	through	through	ADP
cana-6094	15	16	their	their	PRON
cana-6094	15	17	cayley	cayley	ADJ
cana-6094	15	18	graphs	graph	NOUN
cana-6094	15	19	;	;	PUNCT
cana-6094	15	20	topological	topological	ADJ
cana-6094	15	21	spaces	space	NOUN
cana-6094	15	22	through	through	ADP
cana-6094	15	23	their	their	PRON
cana-6094	15	24	simplicial	simplicial	ADJ
cana-6094	15	25	complexes	complex	NOUN
cana-6094	15	26	;	;	PUNCT
cana-6094	15	27	and	and	CCONJ
cana-6094	15	28	combinatorial	combinatorial	ADJ
cana-6094	15	29	problems	problem	NOUN
cana-6094	15	30	through	through	ADP
cana-6094	15	31	their	their	PRON
cana-6094	15	32	constraint	constraint	NOUN
cana-6094	15	33	graphs	graph	NOUN
cana-6094	15	34	.	.	PUNCT
cana-6094	16	1	gnns	gnns	PROPN
cana-6094	16	2	provide	provide	VERB
cana-6094	16	3	a	a	DET
cana-6094	16	4	framework	framework	NOUN
cana-6094	16	5	for	for	ADP
cana-6094	16	6	learning	learn	VERB
cana-6094	16	7	from	from	ADP
cana-6094	16	8	these	these	DET
cana-6094	16	9	representations	representation	NOUN
cana-6094	16	10	by	by	ADP
cana-6094	16	11	propagating	propagate	VERB
cana-6094	16	12	information	information	NOUN
cana-6094	16	13	along	along	ADP
cana-6094	16	14	edges	edge	NOUN
cana-6094	16	15	and	and	CCONJ
cana-6094	16	16	aggregating	aggregate	VERB
cana-6094	16	17	neighborhood	neighborhood	NOUN
cana-6094	16	18	information	information	NOUN
cana-6094	16	19	,	,	PUNCT
cana-6094	16	20	effectively	effectively	ADV
cana-6094	16	21	mimicking	mimic	VERB
cana-6094	16	22	how	how	SCONJ
cana-6094	16	23	mathematicians	mathematician	NOUN
cana-6094	16	24	might	might	AUX
cana-6094	16	25	reason	reason	VERB
cana-6094	16	26	about	about	ADP
cana-6094	16	27	local	local	ADJ
cana-6094	16	28	properties	property	NOUN
cana-6094	16	29	to	to	PART
cana-6094	16	30	deduce	deduce	VERB
cana-6094	16	31	global	global	ADJ
cana-6094	16	32	structure	structure	NOUN
cana-6094	16	33	17	17	NUM
cana-6094	16	34	.	.	PUNCT
cana-6094	17	1	the	the	DET
cana-6094	17	2	fundamental	fundamental	ADJ
cana-6094	17	3	operation	operation	NOUN
cana-6094	17	4	of	of	ADP
cana-6094	17	5	gnns	gnns	ADJ
cana-6094	17	6	message	message	NOUN
cana-6094	17	7	passing	pass	VERB
cana-6094	17	8	enables	enable	VERB
cana-6094	17	9	these	these	DET
cana-6094	17	10	models	model	NOUN
cana-6094	17	11	to	to	PART
cana-6094	17	12	capture	capture	VERB
cana-6094	17	13	the	the	DET
cana-6094	17	14	structural	structural	ADJ
cana-6094	17	15	dependencies	dependency	NOUN
cana-6094	17	16	and	and	CCONJ
cana-6094	17	17	relational	relational	ADJ
cana-6094	17	18	patterns	pattern	NOUN
cana-6094	17	19	inherent	inherent	ADJ
cana-6094	17	20	in	in	ADP
cana-6094	17	21	mathematical	mathematical	ADJ
cana-6094	17	22	objects	object	NOUN
cana-6094	17	23	.	.	PUNCT
cana-6094	18	1	unlike	unlike	ADP
cana-6094	18	2	traditional	traditional	ADJ
cana-6094	18	3	neural	neural	ADJ
cana-6094	18	4	networks	network	NOUN
cana-6094	18	5	designed	design	VERB
cana-6094	18	6	for	for	ADP
cana-6094	18	7	grid	grid	NOUN
cana-6094	18	8	-	-	PUNCT
cana-6094	18	9	like	like	ADJ
cana-6094	18	10	data	datum	NOUN
cana-6094	18	11	(	(	PUNCT
cana-6094	18	12	e.g.	e.g.	ADV
cana-6094	18	13	,	,	PUNCT
cana-6094	18	14	images	image	NOUN
cana-6094	18	15	)	)	PUNCT
cana-6094	18	16	or	or	CCONJ
cana-6094	18	17	sequences	sequence	NOUN
cana-6094	18	18	(	(	PUNCT
cana-6094	18	19	e.g.	e.g.	ADV
cana-6094	18	20	,	,	PUNCT
cana-6094	18	21	text	text	NOUN
cana-6094	18	22	)	)	PUNCT
cana-6094	18	23	,	,	PUNCT
cana-6094	18	24	gnns	gnns	PROPN
cana-6094	18	25	operate	operate	VERB
cana-6094	18	26	directly	directly	ADV
cana-6094	18	27	on	on	ADP
cana-6094	18	28	graph	graph	NOUN
cana-6094	18	29	structures	structure	NOUN
cana-6094	18	30	,	,	PUNCT
cana-6094	18	31	making	make	VERB
cana-6094	18	32	them	they	PRON
cana-6094	18	33	particularly	particularly	ADV
cana-6094	18	34	suited	suit	VERB
cana-6094	18	35	for	for	ADP
cana-6094	18	36	mathematical	mathematical	ADJ
cana-6094	18	37	applications	application	NOUN
cana-6094	18	38	where	where	SCONJ
cana-6094	18	39	relationships	relationship	NOUN
cana-6094	18	40	between	between	ADP
cana-6094	18	41	elements	element	NOUN
cana-6094	18	42	are	be	AUX
cana-6094	18	43	as	as	ADV
cana-6094	18	44	important	important	ADJ
cana-6094	18	45	as	as	ADP
cana-6094	18	46	the	the	DET
cana-6094	18	47	elements	element	NOUN
cana-6094	18	48	themselves	themselves	PRON
cana-6094	18	49	47	47	NUM
cana-6094	18	50	.	.	PUNCT
cana-6094	19	1	this	this	DET
cana-6094	19	2	capability	capability	NOUN
cana-6094	19	3	has	have	AUX
cana-6094	19	4	led	lead	VERB
cana-6094	19	5	to	to	ADP
cana-6094	19	6	successful	successful	ADJ
cana-6094	19	7	applications	application	NOUN
cana-6094	19	8	across	across	ADP
cana-6094	19	9	diverse	diverse	ADJ
cana-6094	19	10	mathematical	mathematical	ADJ
cana-6094	19	11	domains	domain	NOUN
cana-6094	19	12	,	,	PUNCT
cana-6094	19	13	from	from	ADP
cana-6094	19	14	identifying	identify	VERB
cana-6094	19	15	group	group	NOUN
cana-6094	19	16	properties	property	NOUN
cana-6094	19	17	and	and	CCONJ
cana-6094	19	18	characterizing	characterize	VERB
cana-6094	19	19	topological	topological	ADJ
cana-6094	19	20	invariants	invariant	NOUN
cana-6094	19	21	to	to	ADP
cana-6094	19	22	solving	solve	VERB
cana-6094	19	23	combinatorial	combinatorial	ADJ
cana-6094	19	24	optimization	optimization	NOUN
cana-6094	19	25	problems	problem	NOUN
cana-6094	19	26	.	.	PUNCT
cana-6094	20	1	this	this	DET
cana-6094	20	2	paper	paper	NOUN
cana-6094	20	3	provides	provide	VERB
cana-6094	20	4	a	a	DET
cana-6094	20	5	comprehensive	comprehensive	ADJ
cana-6094	20	6	examination	examination	NOUN
cana-6094	20	7	of	of	ADP
cana-6094	20	8	gnn	gnn	PROPN
cana-6094	20	9	methodologies	methodology	NOUN
cana-6094	20	10	for	for	ADP
cana-6094	20	11	analyzing	analyze	VERB
cana-6094	20	12	complex	complex	ADJ
cana-6094	20	13	mathematical	mathematical	ADJ
cana-6094	20	14	structures	structure	NOUN
cana-6094	20	15	.	.	PUNCT
cana-6094	21	1	we	we	PRON
cana-6094	21	2	review	review	VERB
cana-6094	21	3	theoretical	theoretical	ADJ
cana-6094	21	4	foundations	foundation	NOUN
cana-6094	21	5	,	,	PUNCT
cana-6094	21	6	architectural	architectural	ADJ
cana-6094	21	7	variants	variant	NOUN
cana-6094	21	8	,	,	PUNCT
cana-6094	21	9	and	and	CCONJ
cana-6094	21	10	application	application	NOUN
cana-6094	21	11	domains	domain	NOUN
cana-6094	21	12	,	,	PUNCT
cana-6094	21	13	highlighting	highlight	VERB
cana-6094	21	14	how	how	SCONJ
cana-6094	21	15	gnns	gnns	NOUN
cana-6094	21	16	can	can	AUX
cana-6094	21	17	advance	advance	VERB
cana-6094	21	18	mathematical	mathematical	ADJ
cana-6094	21	19	discovery	discovery	NOUN
cana-6094	21	20	through	through	ADP
cana-6094	21	21	data	data	NOUN
cana-6094	21	22	-	-	PUNCT
cana-6094	21	23	driven	drive	VERB
cana-6094	21	24	pattern	pattern	NOUN
cana-6094	21	25	recognition	recognition	NOUN
cana-6094	21	26	complemented	complement	VERB
cana-6094	21	27	by	by	ADP
cana-6094	21	28	formal	formal	ADJ
cana-6094	21	29	reasoning	reasoning	NOUN
cana-6094	21	30	.	.	PUNCT
cana-6094	22	1	through	through	ADP
cana-6094	22	2	this	this	DET
cana-6094	22	3	synthesis	synthesis	NOUN
cana-6094	22	4	,	,	PUNCT
cana-6094	22	5	we	we	PRON
cana-6094	22	6	aim	aim	VERB
cana-6094	22	7	to	to	PART
cana-6094	22	8	bridge	bridge	VERB
cana-6094	22	9	the	the	DET
cana-6094	22	10	gap	gap	NOUN
cana-6094	22	11	between	between	ADP
cana-6094	22	12	graph	graph	NOUN
cana-6094	22	13	representation	representation	NOUN
cana-6094	22	14	learning	learning	NOUN
cana-6094	22	15	and	and	CCONJ
cana-6094	22	16	mathematical	mathematical	ADJ
cana-6094	22	17	analysis	analysis	NOUN
cana-6094	22	18	,	,	PUNCT
cana-6094	22	19	providing	provide	VERB
cana-6094	22	20	researchers	researcher	NOUN
cana-6094	22	21	with	with	ADP
cana-6094	22	22	a	a	DET
cana-6094	22	23	foundation	foundation	NOUN
cana-6094	22	24	for	for	ADP
cana-6094	22	25	leveraging	leverage	VERB
cana-6094	22	26	gnns	gnn	NOUN
cana-6094	22	27	in	in	ADP
cana-6094	22	28	their	their	PRON
cana-6094	22	29	mathematical	mathematical	ADJ
cana-6094	22	30	explorations	exploration	NOUN
cana-6094	22	31	while	while	SCONJ
cana-6094	22	32	identifying	identify	VERB
cana-6094	22	33	promising	promising	ADJ
cana-6094	22	34	directions	direction	NOUN
cana-6094	22	35	for	for	ADP
cana-6094	22	36	future	future	ADJ
cana-6094	22	37	research	research	NOUN
cana-6094	22	38	.	.	PUNCT
cana-6094	23	1	2.theoretical	2.theoretical	NUM
cana-6094	23	2	foundations	foundation	NOUN
cana-6094	23	3	2.1	2.1	NUM
cana-6094	23	4	graph	graph	NOUN
cana-6094	23	5	theory	theory	NOUN
cana-6094	23	6	and	and	CCONJ
cana-6094	23	7	mathematical	mathematical	ADJ
cana-6094	23	8	structures	structure	NOUN
cana-6094	23	9	graph	graph	NOUN
cana-6094	23	10	theory	theory	NOUN
cana-6094	23	11	offers	offer	VERB
cana-6094	23	12	a	a	DET
cana-6094	23	13	robust	robust	ADJ
cana-6094	23	14	mathematical	mathematical	ADJ
cana-6094	23	15	framework	framework	NOUN
cana-6094	23	16	for	for	ADP
cana-6094	23	17	understanding	understanding	NOUN
cana-6094	23	18	and	and	CCONJ
cana-6094	23	19	analyzing	analyze	VERB
cana-6094	23	20	relational	relational	ADJ
cana-6094	23	21	structures	structure	NOUN
cana-6094	23	22	,	,	PUNCT
cana-6094	23	23	making	make	VERB
cana-6094	23	24	it	it	PRON
cana-6094	23	25	inherently	inherently	ADV
cana-6094	23	26	well	well	ADV
cana-6094	23	27	-	-	PUNCT
cana-6094	23	28	suited	suited	ADJ
cana-6094	23	29	for	for	ADP
cana-6094	23	30	representing	represent	VERB
cana-6094	23	31	complex	complex	ADJ
cana-6094	23	32	mathematical	mathematical	ADJ
cana-6094	23	33	objects	object	NOUN
cana-6094	23	34	.	.	PUNCT
cana-6094	24	1	a	a	DET
cana-6094	24	2	graph	graph	NOUN
cana-6094	24	3	g=(v	g=(v	NOUN
cana-6094	24	4	,	,	PUNCT
cana-6094	24	5	e	e	NOUN
cana-6094	24	6	)	)	PUNCT
cana-6094	24	7	comprises	comprise	VERB
cana-6094	24	8	a	a	DET
cana-6094	24	9	set	set	NOUN
cana-6094	24	10	of	of	ADP
cana-6094	24	11	vertices	vertex	NOUN
cana-6094	24	12	(	(	PUNCT
cana-6094	24	13	or	or	CCONJ
cana-6094	24	14	nodes	node	NOUN
cana-6094	24	15	)	)	PUNCT
cana-6094	24	16	v	v	NOUN
cana-6094	24	17	and	and	CCONJ
cana-6094	24	18	a	a	DET
cana-6094	24	19	set	set	NOUN
cana-6094	24	20	of	of	ADP
cana-6094	24	21	edges	edge	NOUN
cana-6094	24	22	e	e	NOUN
cana-6094	24	23	that	that	PRON
cana-6094	24	24	connect	connect	VERB
cana-6094	24	25	pairs	pair	NOUN
cana-6094	24	26	of	of	ADP
cana-6094	24	27	vertices	vertex	NOUN
cana-6094	24	28	.	.	PUNCT
cana-6094	25	1	in	in	ADP
cana-6094	25	2	mathematical	mathematical	ADJ
cana-6094	25	3	contexts	contexts	NOUN
cana-6094	25	4	,	,	PUNCT
cana-6094	25	5	vertices	vertice	VERB
cana-6094	25	6	typically	typically	ADV
cana-6094	25	7	correspond	correspond	VERB
cana-6094	25	8	to	to	ADP
cana-6094	25	9	fundamental	fundamental	ADJ
cana-6094	25	10	elements	element	NOUN
cana-6094	25	11	such	such	ADJ
cana-6094	25	12	as	as	ADP
cana-6094	25	13	group	group	NOUN
cana-6094	25	14	elements	element	NOUN
cana-6094	25	15	,	,	PUNCT
cana-6094	25	16	topological	topological	ADJ
cana-6094	25	17	points	point	NOUN
cana-6094	25	18	,	,	PUNCT
cana-6094	25	19	or	or	CCONJ
cana-6094	25	20	combinatorial	combinatorial	ADJ
cana-6094	25	21	objects	object	NOUN
cana-6094	25	22	while	while	SCONJ
cana-6094	25	23	edges	edge	NOUN
cana-6094	25	24	represent	represent	VERB
cana-6094	25	25	the	the	DET
cana-6094	25	26	relationships	relationship	NOUN
cana-6094	25	27	between	between	ADP
cana-6094	25	28	these	these	DET
cana-6094	25	29	elements	element	NOUN
cana-6094	25	30	,	,	PUNCT
cana-6094	25	31	including	include	VERB
cana-6094	25	32	group	group	NOUN
cana-6094	25	33	operations	operation	NOUN
cana-6094	25	34	,	,	PUNCT
cana-6094	25	35	topological	topological	ADJ
cana-6094	25	36	connections	connection	NOUN
cana-6094	25	37	,	,	PUNCT
cana-6094	25	38	or	or	CCONJ
cana-6094	25	39	combinatorial	combinatorial	ADJ
cana-6094	25	40	constraints	constraint	NOUN
cana-6094	25	41	..	..	PUNCT
cana-6094	26	1	many	many	ADJ
cana-6094	26	2	mathematical	mathematical	ADJ
cana-6094	26	3	structures	structure	NOUN
cana-6094	26	4	admit	admit	VERB
cana-6094	26	5	natural	natural	ADJ
cana-6094	26	6	graph	graph	NOUN
cana-6094	26	7	representations	representation	NOUN
cana-6094	26	8	:	:	PUNCT
cana-6094	26	9	•	•	ADP
cana-6094	26	10	algebraic	algebraic	ADJ
cana-6094	26	11	structures	structure	NOUN
cana-6094	26	12	:	:	PUNCT
cana-6094	26	13	groups	group	NOUN
cana-6094	26	14	can	can	AUX
cana-6094	26	15	be	be	AUX
cana-6094	26	16	represented	represent	VERB
cana-6094	26	17	by	by	ADP
cana-6094	26	18	their	their	PRON
cana-6094	26	19	cayley	cayley	ADJ
cana-6094	26	20	graphs	graph	NOUN
cana-6094	26	21	where	where	SCONJ
cana-6094	26	22	vertices	vertex	NOUN
cana-6094	26	23	correspond	correspond	VERB
cana-6094	26	24	to	to	ADP
cana-6094	26	25	group	group	NOUN
cana-6094	26	26	elements	element	NOUN
cana-6094	26	27	and	and	CCONJ
cana-6094	26	28	edges	edge	NOUN
cana-6094	26	29	represent	represent	VERB
cana-6094	26	30	generator	generator	NOUN
cana-6094	26	31	applications	application	NOUN
cana-6094	26	32	.	.	PUNCT
cana-6094	27	1	rings	ring	NOUN
cana-6094	27	2	and	and	CCONJ
cana-6094	27	3	fields	field	NOUN
cana-6094	27	4	can	can	AUX
cana-6094	27	5	be	be	AUX
cana-6094	27	6	represented	represent	VERB
cana-6094	27	7	through	through	ADP
cana-6094	27	8	their	their	PRON
cana-6094	27	9	multiplicative	multiplicative	ADJ
cana-6094	27	10	and	and	CCONJ
cana-6094	27	11	additive	additive	ADJ
cana-6094	27	12	structures	structure	NOUN
cana-6094	27	13	7	7	NUM
cana-6094	27	14	.	.	NOUN
cana-6094	27	15	•	•	NUM
cana-6094	27	16	geometric	geometric	ADJ
cana-6094	27	17	structures	structure	NOUN
cana-6094	27	18	:	:	PUNCT
cana-6094	27	19	topological	topological	ADJ
cana-6094	27	20	spaces	space	NOUN
cana-6094	27	21	can	can	AUX
cana-6094	27	22	be	be	AUX
cana-6094	27	23	represented	represent	VERB
cana-6094	27	24	as	as	ADP
cana-6094	27	25	simplicial	simplicial	ADJ
cana-6094	27	26	complexes	complex	NOUN
cana-6094	27	27	or	or	CCONJ
cana-6094	27	28	cell	cell	NOUN
cana-6094	27	29	complexes	complex	NOUN
cana-6094	27	30	,	,	PUNCT
cana-6094	27	31	which	which	PRON
cana-6094	27	32	are	be	AUX
cana-6094	27	33	hypergraphs	hypergraph	NOUN
cana-6094	27	34	capturing	capture	VERB
cana-6094	27	35	connectivity	connectivity	NOUN
cana-6094	27	36	relationships	relationship	NOUN
cana-6094	27	37	.	.	PUNCT
cana-6094	28	1	manifolds	manifold	NOUN
cana-6094	28	2	can	can	AUX
cana-6094	28	3	be	be	AUX
cana-6094	28	4	discretized	discretize	VERB
cana-6094	28	5	as	as	ADP
cana-6094	28	6	graphs	graph	NOUN
cana-6094	28	7	while	while	SCONJ
cana-6094	28	8	preserving	preserve	VERB
cana-6094	28	9	their	their	PRON
cana-6094	28	10	geometric	geometric	ADJ
cana-6094	28	11	properties	property	NOUN
cana-6094	28	12	8	8	NUM
cana-6094	28	13	.	.	NOUN
cana-6094	28	14	•	•	NUM
cana-6094	28	15	combinatorial	combinatorial	ADJ
cana-6094	28	16	structures	structure	NOUN
cana-6094	28	17	:	:	PUNCT
cana-6094	28	18	matroids	matroid	NOUN
cana-6094	28	19	,	,	PUNCT
cana-6094	28	20	designs	design	NOUN
cana-6094	28	21	,	,	PUNCT
cana-6094	28	22	and	and	CCONJ
cana-6094	28	23	other	other	ADJ
cana-6094	28	24	combinatorial	combinatorial	ADJ
cana-6094	28	25	objects	object	NOUN
cana-6094	28	26	can	can	AUX
cana-6094	28	27	be	be	AUX
cana-6094	28	28	represented	represent	VERB
cana-6094	28	29	as	as	ADP
cana-6094	28	30	bipartite	bipartite	NOUN
cana-6094	28	31	graphs	graph	NOUN
cana-6094	28	32	or	or	CCONJ
cana-6094	28	33	incidence	incidence	NOUN
cana-6094	28	34	structures	structure	NOUN
cana-6094	28	35	that	that	PRON
cana-6094	28	36	capture	capture	VERB
cana-6094	28	37	their	their	PRON
cana-6094	28	38	combinatorial	combinatorial	ADJ
cana-6094	28	39	relationships	relationship	NOUN
cana-6094	28	40	1	1	NUM
cana-6094	28	41	.	.	PUNCT
cana-6094	29	1	these	these	DET
cana-6094	29	2	graph	graph	NOUN
cana-6094	29	3	representations	representation	NOUN
cana-6094	29	4	provide	provide	VERB
cana-6094	29	5	the	the	DET
cana-6094	29	6	foundation	foundation	NOUN
cana-6094	29	7	for	for	ADP
cana-6094	29	8	applying	apply	VERB
cana-6094	29	9	gnns	gnn	NOUN
cana-6094	29	10	to	to	ADP
cana-6094	29	11	mathematical	mathematical	ADJ
cana-6094	29	12	analysis	analysis	NOUN
cana-6094	29	13	,	,	PUNCT
cana-6094	29	14	enabling	enable	VERB
cana-6094	29	15	these	these	DET
cana-6094	29	16	models	model	NOUN
cana-6094	29	17	to	to	PART
cana-6094	29	18	learn	learn	VERB
cana-6094	29	19	from	from	ADP
cana-6094	29	20	the	the	DET
cana-6094	29	21	structural	structural	ADJ
cana-6094	29	22	properties	property	NOUN
cana-6094	29	23	of	of	ADP
cana-6094	29	24	mathematical	mathematical	ADJ
cana-6094	29	25	objects	object	NOUN
cana-6094	29	26	.	.	PUNCT
cana-6094	30	1	i.	i.	NOUN
cana-6094	30	2	2.2	2.2	NUM
cana-6094	30	3	graph	graph	NOUN
cana-6094	30	4	neural	neural	ADJ
cana-6094	30	5	networks	network	NOUN
cana-6094	30	6	fundamentals	fundamental	NOUN
cana-6094	30	7	graph	graph	VERB
cana-6094	30	8	neural	neural	ADJ
cana-6094	30	9	networks	network	NOUN
cana-6094	30	10	(	(	PUNCT
cana-6094	30	11	gnns	gnns	NOUN
cana-6094	30	12	)	)	PUNCT
cana-6094	30	13	are	be	AUX
cana-6094	30	14	a	a	DET
cana-6094	30	15	class	class	NOUN
cana-6094	30	16	of	of	ADP
cana-6094	30	17	deep	deep	ADJ
cana-6094	30	18	learning	learning	NOUN
cana-6094	30	19	models	model	NOUN
cana-6094	30	20	designed	design	VERB
cana-6094	30	21	to	to	PART
cana-6094	30	22	process	process	VERB
cana-6094	30	23	graph	graph	NOUN
cana-6094	30	24	-	-	PUNCT
cana-6094	30	25	structured	structure	VERB
cana-6094	30	26	data	datum	NOUN
cana-6094	30	27	.	.	PUNCT
cana-6094	31	1	unlike	unlike	ADP
cana-6094	31	2	traditional	traditional	ADJ
cana-6094	31	3	neural	neural	ADJ
cana-6094	31	4	networks	network	NOUN
cana-6094	31	5	that	that	PRON
cana-6094	31	6	assume	assume	VERB
cana-6094	31	7	euclidean	euclidean	ADJ
cana-6094	31	8	grid	grid	NOUN
cana-6094	31	9	structures	structure	NOUN
cana-6094	31	10	,	,	PUNCT
cana-6094	31	11	gnns	gnn	NOUN
cana-6094	31	12	operate	operate	VERB
cana-6094	31	13	directly	directly	ADV
cana-6094	31	14	on	on	ADP
cana-6094	31	15	graphs	graph	NOUN
cana-6094	31	16	,	,	PUNCT
cana-6094	31	17	preserving	preserve	VERB
cana-6094	31	18	and	and	CCONJ
cana-6094	31	19	leveraging	leverage	VERB
cana-6094	31	20	their	their	PRON
cana-6094	31	21	relational	relational	ADJ
cana-6094	31	22	structure	structure	NOUN
cana-6094	31	23	14	14	NUM
cana-6094	31	24	.	.	PUNCT
cana-6094	32	1	the	the	DET
cana-6094	32	2	core	core	NOUN
cana-6094	32	3	operation	operation	NOUN
cana-6094	32	4	of	of	ADP
cana-6094	32	5	gnns	gnns	NOUN
cana-6094	32	6	is	be	AUX
cana-6094	32	7	message	message	NOUN
cana-6094	32	8	passing	passing	NOUN
cana-6094	32	9	,	,	PUNCT
cana-6094	32	10	where	where	SCONJ
cana-6094	32	11	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	32	12	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	SCONJ
cana-6094	32	13	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	NOUN
cana-6094	32	14	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	X
cana-6094	32	15	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	X
cana-6094	32	16	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	SYM
cana-6094	32	17	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	32	18	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	32	19	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	NOUN
cana-6094	32	20	communications	communication	NOUN
cana-6094	32	21	on	on	ADP
cana-6094	32	22	applied	apply	VERB
cana-6094	32	23	nonlinear	nonlinear	ADJ
cana-6094	32	24	analysis	analysis	NOUN
cana-6094	32	25	issn	issn	NOUN
cana-6094	32	26	:	:	PUNCT
cana-6094	32	27	1074	1074	NUM
cana-6094	32	28	-	-	PUNCT
cana-6094	32	29	133x	133x	NUM
cana-6094	32	30	vol	vol	NOUN
cana-6094	32	31	29	29	NUM
cana-6094	32	32	no	no	NOUN
cana-6094	32	33	.	.	NOUN
cana-6094	32	34	4	4	NUM
cana-6094	32	35	(	(	PUNCT
cana-6094	32	36	2022	2022	NUM
cana-6094	32	37	)	)	PUNCT
cana-6094	32	38	29	29	NUM
cana-6094	32	39	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	32	40	information	information	NOUN
cana-6094	32	41	is	be	AUX
cana-6094	32	42	propagated	propagate	VERB
cana-6094	32	43	along	along	ADP
cana-6094	32	44	edges	edge	NOUN
cana-6094	32	45	and	and	CCONJ
cana-6094	32	46	aggregated	aggregate	VERB
cana-6094	32	47	at	at	ADP
cana-6094	32	48	nodes	node	NOUN
cana-6094	32	49	to	to	PART
cana-6094	32	50	compute	compute	VERB
cana-6094	32	51	node	node	ADJ
cana-6094	32	52	representations	representation	NOUN
cana-6094	32	53	that	that	PRON
cana-6094	32	54	capture	capture	VERB
cana-6094	32	55	both	both	CCONJ
cana-6094	32	56	local	local	ADJ
cana-6094	32	57	and	and	CCONJ
cana-6094	32	58	global	global	ADJ
cana-6094	32	59	structural	structural	ADJ
cana-6094	32	60	information	information	NOUN
cana-6094	32	61	.	.	PUNCT
cana-6094	33	1	the	the	DET
cana-6094	33	2	message	message	NOUN
cana-6094	33	3	passing	pass	VERB
cana-6094	33	4	framework	framework	NOUN
cana-6094	33	5	can	can	AUX
cana-6094	33	6	be	be	AUX
cana-6094	33	7	formalized	formalize	VERB
cana-6094	33	8	as	as	SCONJ
cana-6094	33	9	follows	follow	VERB
cana-6094	33	10	:	:	PUNCT
cana-6094	33	11	given	give	VERB
cana-6094	33	12	a	a	DET
cana-6094	33	13	graph	graph	NOUN
cana-6094	33	14	g(v	g(v	NOUN
cana-6094	33	15	,	,	PUNCT
cana-6094	33	16	e	e	NOUN
cana-6094	33	17	)	)	PUNCT
cana-6094	33	18	with	with	ADP
cana-6094	33	19	node	node	ADJ
cana-6094	33	20	features	feature	NOUN
cana-6094	33	21	xi	xi	ADP
cana-6094	33	22	for	for	ADP
cana-6094	33	23	each	each	DET
cana-6094	33	24	node	node	NOUN
cana-6094	33	25	i∈v	i∈v	NOUN
cana-6094	33	26	,	,	PUNCT
cana-6094	33	27	a	a	DET
cana-6094	33	28	graph	graph	NOUN
cana-6094	33	29	neural	neural	ADJ
cana-6094	33	30	network	network	NOUN
cana-6094	33	31	(	(	PUNCT
cana-6094	33	32	gnn	gnn	NOUN
cana-6094	33	33	)	)	PUNCT
cana-6094	33	34	layer	layer	NOUN
cana-6094	33	35	computes	compute	VERB
cana-6094	33	36	updated	update	VERB
cana-6094	33	37	node	node	ADJ
cana-6094	33	38	representations	representation	NOUN
cana-6094	33	39	hi	hi	INTJ
cana-6094	33	40	through	through	ADP
cana-6094	33	41	three	three	NUM
cana-6094	33	42	fundamental	fundamental	ADJ
cana-6094	33	43	operations	operation	NOUN
cana-6094	33	44	:	:	PUNCT
cana-6094	33	45	message	message	NOUN
cana-6094	33	46	generation	generation	NOUN
cana-6094	33	47	,	,	PUNCT
cana-6094	33	48	message	message	NOUN
cana-6094	33	49	aggregation	aggregation	NOUN
cana-6094	33	50	,	,	PUNCT
cana-6094	33	51	and	and	CCONJ
cana-6094	33	52	node	node	ADJ
cana-6094	33	53	update	update	NOUN
cana-6094	33	54	.	.	PUNCT
cana-6094	34	1	1	1	X
cana-6094	34	2	.	.	X
cana-6094	34	3	message	message	NOUN
cana-6094	34	4	computation	computation	NOUN
cana-6094	34	5	:	:	PUNCT
cana-6094	34	6	for	for	ADP
cana-6094	34	7	each	each	DET
cana-6094	34	8	edge	edge	NOUN
cana-6094	34	9	(	(	PUNCT
cana-6094	34	10	i	i	PROPN
cana-6094	34	11	,	,	PUNCT
cana-6094	34	12	j	j	PROPN
cana-6094	34	13	)	)	PUNCT
cana-6094	34	14	,	,	PUNCT
cana-6094	34	15	compute	compute	VERB
cana-6094	34	16	a	a	DET
cana-6094	34	17	message	message	NOUN
cana-6094	34	18	m_{ij	m_{ij	NOUN
cana-6094	34	19	}	}	PUNCT
cana-6094	34	20	=	=	SYM
cana-6094	34	21	f(x_i	f(x_i	PROPN
cana-6094	34	22	,	,	PUNCT
cana-6094	34	23	x_j	x_j	NUM
cana-6094	34	24	,	,	PUNCT
cana-6094	34	25	e_{ij	e_{ij	NOUN
cana-6094	34	26	}	}	PUNCT
cana-6094	34	27	)	)	PUNCT
cana-6094	34	28	,	,	PUNCT
cana-6094	34	29	where	where	SCONJ
cana-6094	34	30	f	f	PROPN
cana-6094	34	31	is	be	AUX
cana-6094	34	32	a	a	DET
cana-6094	34	33	message	message	NOUN
cana-6094	34	34	function	function	NOUN
cana-6094	34	35	(	(	PUNCT
cana-6094	34	36	typically	typically	ADV
cana-6094	34	37	a	a	DET
cana-6094	34	38	neural	neural	ADJ
cana-6094	34	39	network	network	NOUN
cana-6094	34	40	)	)	PUNCT
cana-6094	34	41	and	and	CCONJ
cana-6094	34	42	e_{ij	e_{ij	NOUN
cana-6094	34	43	}	}	PUNCT
cana-6094	34	44	represents	represent	VERB
cana-6094	34	45	optional	optional	ADJ
cana-6094	34	46	edge	edge	NOUN
cana-6094	34	47	features	feature	NOUN
cana-6094	34	48	.	.	PUNCT
cana-6094	35	1	2	2	X
cana-6094	35	2	.	.	X
cana-6094	35	3	aggregation	aggregation	NOUN
cana-6094	35	4	:	:	PUNCT
cana-6094	35	5	for	for	ADP
cana-6094	35	6	each	each	DET
cana-6094	35	7	node	node	PROPN
cana-6094	35	8	ii	ii	PROPN
cana-6094	35	9	,	,	PUNCT
cana-6094	35	10	messages	message	NOUN
cana-6094	35	11	from	from	ADP
cana-6094	35	12	its	its	PRON
cana-6094	35	13	neighborhood	neighborhood	NOUN
cana-6094	35	14	n(i)\mathcal{n}(i	n(i)\mathcal{n}(i	NOUN
cana-6094	35	15	)	)	PUNCT
cana-6094	35	16	are	be	AUX
cana-6094	35	17	aggregated	aggregate	VERB
cana-6094	35	18	as	as	SCONJ
cana-6094	35	19	follows	follow	VERB
cana-6094	35	20	:	:	PUNCT
cana-6094	35	21	mi	mi	PROPN
cana-6094	35	22	=	=	PROPN
cana-6094	35	23	g({mij	g({mij	PROPN
cana-6094	35	24	:	:	PUNCT
cana-6094	35	25	j∈n(i	j∈n(i	NOUN
cana-6094	35	26	)	)	PUNCT
cana-6094	35	27	}	}	PUNCT
cana-6094	35	28	)	)	PUNCT
cana-6094	35	29	,	,	PUNCT
cana-6094	35	30	3	3	X
cana-6094	35	31	.	.	X
cana-6094	35	32	update	update	NOUN
cana-6094	35	33	:	:	PUNCT
cana-6094	35	34	update	update	VERB
cana-6094	35	35	the	the	DET
cana-6094	35	36	node	node	NOUN
cana-6094	35	37	representation	representation	NOUN
cana-6094	35	38	by	by	ADP
cana-6094	35	39	combining	combine	VERB
cana-6094	35	40	the	the	DET
cana-6094	35	41	aggregated	aggregated	ADJ
cana-6094	35	42	message	message	NOUN
cana-6094	35	43	with	with	ADP
cana-6094	35	44	the	the	DET
cana-6094	35	45	current	current	ADJ
cana-6094	35	46	node	node	ADJ
cana-6094	35	47	state	state	NOUN
cana-6094	35	48	:	:	PUNCT
cana-6094	35	49	h_i	h_i	PUNCT
cana-6094	35	50	=	=	SYM
cana-6094	35	51	k(x_i	k(x_i	PROPN
cana-6094	35	52	,	,	PUNCT
cana-6094	35	53	r{m}_i	r{m}_i	ADJ
cana-6094	35	54	)	)	PUNCT
cana-6094	35	55	,	,	PUNCT
cana-6094	35	56	where	where	SCONJ
cana-6094	35	57	k	k	PROPN
cana-6094	35	58	is	be	AUX
cana-6094	35	59	an	an	DET
cana-6094	35	60	update	update	NOUN
cana-6094	35	61	function	function	NOUN
cana-6094	35	62	(	(	PUNCT
cana-6094	35	63	typically	typically	ADV
cana-6094	35	64	a	a	DET
cana-6094	35	65	neural	neural	ADJ
cana-6094	35	66	network	network	NOUN
cana-6094	35	67	)	)	PUNCT
cana-6094	35	68	.	.	PUNCT
cana-6094	36	1	through	through	ADP
cana-6094	36	2	multiple	multiple	ADJ
cana-6094	36	3	layers	layer	NOUN
cana-6094	36	4	of	of	ADP
cana-6094	36	5	message	message	NOUN
cana-6094	36	6	passing	passing	NOUN
cana-6094	36	7	,	,	PUNCT
cana-6094	36	8	gnns	gnn	NOUN
cana-6094	36	9	can	can	AUX
cana-6094	36	10	capture	capture	VERB
cana-6094	36	11	increasingly	increasingly	ADV
cana-6094	36	12	complex	complex	ADJ
cana-6094	36	13	structural	structural	ADJ
cana-6094	36	14	patterns	pattern	NOUN
cana-6094	36	15	and	and	CCONJ
cana-6094	36	16	dependencies	dependency	NOUN
cana-6094	36	17	,	,	PUNCT
cana-6094	36	18	enabling	enable	VERB
cana-6094	36	19	them	they	PRON
cana-6094	36	20	to	to	PART
cana-6094	36	21	learn	learn	VERB
cana-6094	36	22	rich	rich	ADJ
cana-6094	36	23	representations	representation	NOUN
cana-6094	36	24	of	of	ADP
cana-6094	36	25	mathematical	mathematical	ADJ
cana-6094	36	26	structures	structure	NOUN
cana-6094	36	27	7	7	NUM
cana-6094	36	28	.	.	PUNCT
cana-6094	36	29	mathematical	mathematical	ADJ
cana-6094	36	30	structure	structure	NOUN
cana-6094	36	31	graph	graph	NOUN
cana-6094	36	32	representation	representation	NOUN
cana-6094	36	33	node	node	NOUN
cana-6094	36	34	meaning	mean	VERB
cana-6094	36	35	edge	edge	NOUN
cana-6094	36	36	meaning	mean	VERB
cana-6094	36	37	group	group	NOUN
cana-6094	36	38	cayley	cayley	NOUN
cana-6094	36	39	graph	graph	NOUN
cana-6094	36	40	group	group	NOUN
cana-6094	36	41	element	element	NOUN
cana-6094	36	42	generator	generator	NOUN
cana-6094	36	43	application	application	NOUN
cana-6094	36	44	ring	ring	NOUN
cana-6094	36	45	/	/	SYM
cana-6094	36	46	field	field	NOUN
cana-6094	36	47	bipartite	bipartite	NOUN
cana-6094	36	48	graph	graph	NOUN
cana-6094	36	49	ring	ring	NOUN
cana-6094	36	50	elements	element	NOUN
cana-6094	36	51	operations	operation	NOUN
cana-6094	36	52	between	between	ADP
cana-6094	36	53	elements	element	NOUN
cana-6094	36	54	topological	topological	ADJ
cana-6094	36	55	space	space	NOUN
cana-6094	36	56	simplicial	simplicial	ADJ
cana-6094	36	57	complex	complex	ADJ
cana-6094	36	58	points	point	NOUN
cana-6094	36	59	,	,	PUNCT
cana-6094	36	60	simplices	simplice	NOUN
cana-6094	36	61	face	face	VERB
cana-6094	36	62	relations	relation	NOUN
cana-6094	36	63	manifold	manifold	ADJ
cana-6094	36	64	discretization	discretization	NOUN
cana-6094	36	65	grid	grid	NOUN
cana-6094	36	66	points	point	NOUN
cana-6094	36	67	on	on	ADP
cana-6094	36	68	manifold	manifold	ADJ
cana-6094	36	69	geodesic	geodesic	ADJ
cana-6094	36	70	connections	connection	NOUN
cana-6094	36	71	combinatorial	combinatorial	ADJ
cana-6094	36	72	design	design	NOUN
cana-6094	36	73	incidence	incidence	NOUN
cana-6094	36	74	graph	graph	NOUN
cana-6094	36	75	points	point	NOUN
cana-6094	36	76	and	and	CCONJ
cana-6094	36	77	blocks	block	VERB
cana-6094	36	78	incidence	incidence	NOUN
cana-6094	36	79	relations	relation	NOUN
cana-6094	36	80	graph	graph	VERB
cana-6094	36	81	itself	itself	PRON
cana-6094	36	82	original	original	ADJ
cana-6094	36	83	graph	graph	NOUN
cana-6094	36	84	vertex	vertex	NOUN
cana-6094	36	85	edge	edge	NOUN
cana-6094	36	86	table	table	NOUN
cana-6094	37	1	1	1	NUM
cana-6094	37	2	:	:	PUNCT
cana-6094	37	3	mathematical	mathematical	ADJ
cana-6094	37	4	structures	structure	NOUN
cana-6094	37	5	and	and	CCONJ
cana-6094	37	6	their	their	PRON
cana-6094	37	7	graph	graph	NOUN
cana-6094	37	8	representations	representation	VERB
cana-6094	37	9	2.3	2.3	NUM
cana-6094	37	10	inductive	inductive	ADJ
cana-6094	37	11	biases	bias	NOUN
cana-6094	37	12	for	for	ADP
cana-6094	37	13	mathematical	mathematical	ADJ
cana-6094	37	14	reasoning	reasoning	NOUN
cana-6094	37	15	gnns	gnn	NOUN
cana-6094	37	16	incorporate	incorporate	VERB
cana-6094	37	17	several	several	ADJ
cana-6094	37	18	inductive	inductive	ADJ
cana-6094	37	19	biases	bias	NOUN
cana-6094	37	20	that	that	PRON
cana-6094	37	21	make	make	VERB
cana-6094	37	22	them	they	PRON
cana-6094	37	23	particularly	particularly	ADV
cana-6094	37	24	suited	suit	VERB
cana-6094	37	25	for	for	ADP
cana-6094	37	26	mathematical	mathematical	ADJ
cana-6094	37	27	reasoning	reasoning	NOUN
cana-6094	37	28	:	:	PUNCT
cana-6094	37	29	•	•	NUM
cana-6094	37	30	permutation	permutation	NOUN
cana-6094	37	31	invariance	invariance	NOUN
cana-6094	37	32	:	:	PUNCT
cana-6094	37	33	gnns	gnn	NOUN
cana-6094	37	34	are	be	AUX
cana-6094	37	35	inherently	inherently	ADV
cana-6094	37	36	invariant	invariant	ADJ
cana-6094	37	37	to	to	PART
cana-6094	37	38	node	node	NOUN
cana-6094	37	39	ordering	ordering	NOUN
cana-6094	37	40	,	,	PUNCT
cana-6094	37	41	making	make	VERB
cana-6094	37	42	them	they	PRON
cana-6094	37	43	suitable	suitable	ADJ
cana-6094	37	44	for	for	ADP
cana-6094	37	45	mathematical	mathematical	ADJ
cana-6094	37	46	structures	structure	NOUN
cana-6094	37	47	where	where	SCONJ
cana-6094	37	48	the	the	DET
cana-6094	37	49	labeling	labeling	NOUN
cana-6094	37	50	of	of	ADP
cana-6094	37	51	elements	element	NOUN
cana-6094	37	52	is	be	AUX
cana-6094	37	53	arbitrary	arbitrary	ADJ
cana-6094	37	54	2	2	NUM
cana-6094	37	55	.	.	PUNCT
cana-6094	38	1	this	this	DET
cana-6094	38	2	property	property	NOUN
cana-6094	38	3	ensures	ensure	VERB
cana-6094	38	4	that	that	SCONJ
cana-6094	38	5	the	the	DET
cana-6094	38	6	analysis	analysis	NOUN
cana-6094	38	7	focuses	focus	VERB
cana-6094	38	8	on	on	ADP
cana-6094	38	9	the	the	DET
cana-6094	38	10	intrinsic	intrinsic	ADJ
cana-6094	38	11	structure	structure	NOUN
cana-6094	38	12	rather	rather	ADV
cana-6094	38	13	than	than	ADP
cana-6094	38	14	arbitrary	arbitrary	ADJ
cana-6094	38	15	representations	representation	NOUN
cana-6094	38	16	.	.	PUNCT
cana-6094	39	1	•	•	NUM
cana-6094	39	2	locality	locality	NOUN
cana-6094	39	3	:	:	PUNCT
cana-6094	39	4	the	the	DET
cana-6094	39	5	message	message	NOUN
cana-6094	39	6	passing	pass	VERB
cana-6094	39	7	mechanism	mechanism	NOUN
cana-6094	39	8	emphasizes	emphasize	VERB
cana-6094	39	9	local	local	ADJ
cana-6094	39	10	neighborhoods	neighborhood	NOUN
cana-6094	39	11	,	,	PUNCT
cana-6094	39	12	mirroring	mirror	VERB
cana-6094	39	13	how	how	SCONJ
cana-6094	39	14	mathematical	mathematical	ADJ
cana-6094	39	15	properties	property	NOUN
cana-6094	39	16	often	often	ADV
cana-6094	39	17	propagate	propagate	VERB
cana-6094	39	18	through	through	ADP
cana-6094	39	19	local	local	ADJ
cana-6094	39	20	relations	relation	NOUN
cana-6094	39	21	to	to	PART
cana-6094	39	22	determine	determine	VERB
cana-6094	39	23	global	global	ADJ
cana-6094	39	24	structure	structure	NOUN
cana-6094	39	25	7	7	NUM
cana-6094	39	26	.	.	PUNCT
cana-6094	40	1	this	this	DET
cana-6094	40	2	locality	locality	NOUN
cana-6094	40	3	bias	bias	NOUN
cana-6094	40	4	is	be	AUX
cana-6094	40	5	particularly	particularly	ADV
cana-6094	40	6	valuable	valuable	ADJ
cana-6094	40	7	for	for	ADP
cana-6094	40	8	understanding	understand	VERB
cana-6094	40	9	local	local	ADJ
cana-6094	40	10	-	-	PUNCT
cana-6094	40	11	to	to	ADP
cana-6094	40	12	-	-	PUNCT
cana-6094	40	13	global	global	ADJ
cana-6094	40	14	principles	principle	NOUN
cana-6094	40	15	in	in	ADP
cana-6094	40	16	mathematics	mathematic	NOUN
cana-6094	40	17	.	.	PUNCT
cana-6094	41	1	•	•	NUM
cana-6094	41	2	compositionality	compositionality	NOUN
cana-6094	41	3	:	:	PUNCT
cana-6094	41	4	multi	multi	ADJ
cana-6094	41	5	-	-	ADJ
cana-6094	41	6	layer	layer	ADJ
cana-6094	41	7	gnns	gnn	NOUN
cana-6094	41	8	naturally	naturally	ADV
cana-6094	41	9	capture	capture	VERB
cana-6094	41	10	hierarchical	hierarchical	ADJ
cana-6094	41	11	compositions	composition	NOUN
cana-6094	41	12	of	of	ADP
cana-6094	41	13	mathematical	mathematical	ADJ
cana-6094	41	14	operations	operation	NOUN
cana-6094	41	15	,	,	PUNCT
cana-6094	41	16	enabling	enable	VERB
cana-6094	41	17	them	they	PRON
cana-6094	41	18	to	to	PART
cana-6094	41	19	learn	learn	VERB
cana-6094	41	20	complex	complex	ADJ
cana-6094	41	21	compositions	composition	NOUN
cana-6094	41	22	of	of	ADP
cana-6094	41	23	simpler	simple	ADJ
cana-6094	41	24	mathematical	mathematical	ADJ
cana-6094	41	25	relationships	relationship	NOUN
cana-6094	41	26	17	17	NUM
cana-6094	41	27	.	.	PUNCT
cana-6094	42	1	these	these	DET
cana-6094	42	2	inductive	inductive	ADJ
cana-6094	42	3	biases	bias	NOUN
cana-6094	42	4	allow	allow	VERB
cana-6094	42	5	gnns	gnn	NOUN
cana-6094	42	6	to	to	PART
cana-6094	42	7	learn	learn	VERB
cana-6094	42	8	efficiently	efficiently	ADV
cana-6094	42	9	from	from	ADP
cana-6094	42	10	mathematical	mathematical	ADJ
cana-6094	42	11	structures	structure	NOUN
cana-6094	42	12	,	,	PUNCT
cana-6094	42	13	often	often	ADV
cana-6094	42	14	generalizing	generalize	VERB
cana-6094	42	15	better	well	ADJ
cana-6094	42	16	than	than	ADP
cana-6094	42	17	architectures	architecture	NOUN
cana-6094	42	18	without	without	ADP
cana-6094	42	19	these	these	DET
cana-6094	42	20	built	build	VERB
cana-6094	42	21	-	-	PUNCT
cana-6094	42	22	in	in	ADP
cana-6094	42	23	assumptions	assumption	NOUN
cana-6094	42	24	about	about	ADP
cana-6094	42	25	the	the	DET
cana-6094	42	26	nature	nature	NOUN
cana-6094	42	27	of	of	ADP
cana-6094	42	28	mathematical	mathematical	ADJ
cana-6094	42	29	relationships	relationship	NOUN
cana-6094	42	30	.	.	PUNCT
cana-6094	43	1	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	PROPN
cana-6094	43	2	https://gnn.seas.upenn.edu/aaai-2025/	https://gnn.seas.upenn.edu/aaai-2025/	PROPN
cana-6094	43	3	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	PROPN
cana-6094	43	4	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	43	5	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	PUNCT
cana-6094	43	6	communications	communication	NOUN
cana-6094	43	7	on	on	ADP
cana-6094	43	8	applied	apply	VERB
cana-6094	43	9	nonlinear	nonlinear	ADJ
cana-6094	43	10	analysis	analysis	NOUN
cana-6094	43	11	issn	issn	NOUN
cana-6094	43	12	:	:	PUNCT
cana-6094	43	13	1074	1074	NUM
cana-6094	43	14	-	-	PUNCT
cana-6094	43	15	133x	133x	NUM
cana-6094	43	16	vol	vol	NOUN
cana-6094	43	17	29	29	NUM
cana-6094	43	18	no	no	NOUN
cana-6094	43	19	.	.	NOUN
cana-6094	43	20	4	4	NUM
cana-6094	43	21	(	(	PUNCT
cana-6094	43	22	2022	2022	NUM
cana-6094	43	23	)	)	PUNCT
cana-6094	43	24	30	30	NUM
cana-6094	43	25	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	43	26	3	3	NUM
cana-6094	43	27	.gnn	.gnn	NOUN
cana-6094	43	28	frameworks	framework	NOUN
cana-6094	43	29	for	for	ADP
cana-6094	43	30	mathematical	mathematical	ADJ
cana-6094	43	31	analysis	analysis	NOUN
cana-6094	43	32	3.1	3.1	NUM
cana-6094	43	33	spectral	spectral	ADJ
cana-6094	43	34	graph	graph	NOUN
cana-6094	43	35	neural	neural	ADJ
cana-6094	43	36	networks	network	NOUN
cana-6094	43	37	spectral	spectral	ADJ
cana-6094	43	38	gnns	gnns	PROPN
cana-6094	43	39	approach	approach	NOUN
cana-6094	43	40	graph	graph	NOUN
cana-6094	43	41	convolution	convolution	NOUN
cana-6094	43	42	from	from	ADP
cana-6094	43	43	the	the	DET
cana-6094	43	44	perspective	perspective	NOUN
cana-6094	43	45	of	of	ADP
cana-6094	43	46	spectral	spectral	ADJ
cana-6094	43	47	graph	graph	NOUN
cana-6094	43	48	theory	theory	NOUN
cana-6094	43	49	,	,	PUNCT
cana-6094	43	50	leveraging	leverage	VERB
cana-6094	43	51	the	the	DET
cana-6094	43	52	graph	graph	NOUN
cana-6094	43	53	laplacian	laplacian	NOUN
cana-6094	43	54	and	and	CCONJ
cana-6094	43	55	its	its	PRON
cana-6094	43	56	spectral	spectral	ADJ
cana-6094	43	57	decomposition	decomposition	NOUN
cana-6094	43	58	to	to	PART
cana-6094	43	59	define	define	VERB
cana-6094	43	60	convolution	convolution	NOUN
cana-6094	43	61	operations	operation	NOUN
cana-6094	43	62	in	in	ADP
cana-6094	43	63	the	the	DET
cana-6094	43	64	fourier	fourier	ADJ
cana-6094	43	65	domain	domain	NOUN
cana-6094	43	66	of	of	ADP
cana-6094	43	67	graphs	graph	NOUN
cana-6094	43	68	.	.	PUNCT
cana-6094	44	1	the	the	DET
cana-6094	44	2	graph	graph	NOUN
cana-6094	44	3	laplacian	laplacian	ADJ
cana-6094	44	4	l	l	NOUN
cana-6094	45	1	=	=	PUNCT
cana-6094	45	2	d	d	X
cana-6094	45	3	a	a	X
cana-6094	45	4	(	(	PUNCT
cana-6094	45	5	where	where	SCONJ
cana-6094	45	6	d	d	NOUN
cana-6094	45	7	is	be	AUX
cana-6094	45	8	the	the	DET
cana-6094	45	9	degree	degree	NOUN
cana-6094	45	10	matrix	matrix	NOUN
cana-6094	45	11	and	and	CCONJ
cana-6094	45	12	a	a	PRON
cana-6094	45	13	is	be	AUX
cana-6094	45	14	the	the	DET
cana-6094	45	15	adjacency	adjacency	PROPN
cana-6094	45	16	matrix	matrix	NOUN
cana-6094	45	17	)	)	PUNCT
cana-6094	45	18	captures	capture	VERB
cana-6094	45	19	important	important	ADJ
cana-6094	45	20	structural	structural	ADJ
cana-6094	45	21	properties	property	NOUN
cana-6094	45	22	of	of	ADP
cana-6094	45	23	graphs	graph	NOUN
cana-6094	45	24	,	,	PUNCT
cana-6094	45	25	with	with	ADP
cana-6094	45	26	its	its	PRON
cana-6094	45	27	eigenvalues	eigenvalue	NOUN
cana-6094	45	28	and	and	CCONJ
cana-6094	45	29	eigenvectors	eigenvector	NOUN
cana-6094	45	30	providing	provide	VERB
cana-6094	45	31	a	a	DET
cana-6094	45	32	natural	natural	ADJ
cana-6094	45	33	basis	basis	NOUN
cana-6094	45	34	for	for	ADP
cana-6094	45	35	graph	graph	NOUN
cana-6094	45	36	fourier	fourier	ADJ
cana-6094	45	37	analysis	analysis	NOUN
cana-6094	45	38	6	6	NUM
cana-6094	45	39	.	.	PUNCT
cana-6094	46	1	spectral	spectral	ADJ
cana-6094	46	2	gnns	gnns	PROPN
cana-6094	46	3	define	define	VERB
cana-6094	46	4	convolution	convolution	NOUN
cana-6094	46	5	operations	operation	NOUN
cana-6094	46	6	as	as	ADP
cana-6094	46	7	multiplication	multiplication	NOUN
cana-6094	46	8	in	in	ADP
cana-6094	46	9	the	the	DET
cana-6094	46	10	spectral	spectral	ADJ
cana-6094	46	11	domain	domain	NOUN
cana-6094	46	12	:	:	PUNCT
cana-6094	46	13	ℎ=(λ)h	ℎ=(λ)h	X
cana-6094	46	14	=	=	SYM
cana-6094	46	15	ugθ(λ)utx	ugθ(λ)utx	NOUN
cana-6094	46	16	where	where	SCONJ
cana-6094	46	17	u	u	NOUN
cana-6094	46	18	is	be	AUX
cana-6094	46	19	the	the	DET
cana-6094	46	20	matrix	matrix	NOUN
cana-6094	46	21	of	of	ADP
cana-6094	46	22	eigenvectors	eigenvector	NOUN
cana-6094	46	23	of	of	ADP
cana-6094	46	24	$	$	SYM
cana-6094	46	25	l$	l$	NOUN
cana-6094	46	26	,	,	PUNCT
cana-6094	46	27	$	$	SYM
cana-6094	46	28	\lambda$	\lambda$	NOUN
cana-6094	46	29	is	be	AUX
cana-6094	46	30	the	the	DET
cana-6094	46	31	diagonal	diagonal	ADJ
cana-6094	46	32	matrix	matrix	NOUN
cana-6094	46	33	of	of	ADP
cana-6094	46	34	eigenvalues	eigenvalue	NOUN
cana-6094	46	35	,	,	PUNCT
cana-6094	46	36	and	and	CCONJ
cana-6094	46	37	$	$	SYM
cana-6094	46	38	g_\theta$	g_\theta$	PRON
cana-6094	46	39	is	be	AUX
cana-6094	46	40	a	a	DET
cana-6094	46	41	learnable	learnable	ADJ
cana-6094	46	42	spectral	spectral	ADJ
cana-6094	46	43	filter	filter	NOUN
cana-6094	46	44	parameterized	parameterize	VERB
cana-6094	46	45	by	by	ADP
cana-6094	46	46	$	$	SYM
cana-6094	46	47	\theta$	\theta$	PRON
cana-6094	46	48	6	6	NUM
cana-6094	46	49	.	.	PUNCT
cana-6094	47	1	this	this	DET
cana-6094	47	2	approach	approach	NOUN
cana-6094	47	3	provides	provide	VERB
cana-6094	47	4	a	a	DET
cana-6094	47	5	mathematical	mathematical	ADJ
cana-6094	47	6	foundation	foundation	NOUN
cana-6094	47	7	for	for	ADP
cana-6094	47	8	graph	graph	NOUN
cana-6094	47	9	convolution	convolution	NOUN
cana-6094	47	10	that	that	PRON
cana-6094	47	11	connects	connect	VERB
cana-6094	47	12	to	to	PART
cana-6094	47	13	harmonic	harmonic	VERB
cana-6094	47	14	analysis	analysis	NOUN
cana-6094	47	15	on	on	ADP
cana-6094	47	16	graphs	graph	NOUN
cana-6094	47	17	.	.	PUNCT
cana-6094	48	1	for	for	ADP
cana-6094	48	2	mathematical	mathematical	ADJ
cana-6094	48	3	structures	structure	NOUN
cana-6094	48	4	,	,	PUNCT
cana-6094	48	5	spectral	spectral	ADJ
cana-6094	48	6	gnns	gnn	NOUN
cana-6094	48	7	are	be	AUX
cana-6094	48	8	particularly	particularly	ADV
cana-6094	48	9	valuable	valuable	ADJ
cana-6094	48	10	for	for	ADP
cana-6094	48	11	capturing	capture	VERB
cana-6094	48	12	global	global	ADJ
cana-6094	48	13	symmetries	symmetry	NOUN
cana-6094	48	14	and	and	CCONJ
cana-6094	48	15	invariant	invariant	ADJ
cana-6094	48	16	properties	property	NOUN
cana-6094	48	17	.	.	PUNCT
cana-6094	49	1	the	the	DET
cana-6094	49	2	spectral	spectral	ADJ
cana-6094	49	3	decomposition	decomposition	NOUN
cana-6094	49	4	naturally	naturally	ADV
cana-6094	49	5	reveals	reveal	VERB
cana-6094	49	6	structural	structural	ADJ
cana-6094	49	7	properties	property	NOUN
cana-6094	49	8	such	such	ADJ
cana-6094	49	9	as	as	ADP
cana-6094	49	10	connectivity	connectivity	NOUN
cana-6094	49	11	,	,	PUNCT
cana-6094	49	12	clustering	clustering	NOUN
cana-6094	49	13	,	,	PUNCT
cana-6094	49	14	and	and	CCONJ
cana-6094	49	15	symmetry	symmetry	NOUN
cana-6094	49	16	patterns	pattern	NOUN
cana-6094	49	17	that	that	PRON
cana-6094	49	18	are	be	AUX
cana-6094	49	19	fundamental	fundamental	ADJ
cana-6094	49	20	to	to	AUX
cana-6094	49	21	mathematical	mathematical	ADJ
cana-6094	49	22	analysis	analysis	NOUN
cana-6094	49	23	6	6	NUM
cana-6094	49	24	.	.	PUNCT
cana-6094	49	25	applications	application	NOUN
cana-6094	49	26	include	include	VERB
cana-6094	49	27	characterizing	characterize	VERB
cana-6094	49	28	isospectral	isospectral	ADJ
cana-6094	49	29	graphs	graph	NOUN
cana-6094	49	30	,	,	PUNCT
cana-6094	49	31	analyzing	analyze	VERB
cana-6094	49	32	expansion	expansion	NOUN
cana-6094	49	33	properties	property	NOUN
cana-6094	49	34	,	,	PUNCT
cana-6094	49	35	and	and	CCONJ
cana-6094	49	36	studying	study	VERB
cana-6094	49	37	symmetry	symmetry	NOUN
cana-6094	49	38	groups	group	NOUN
cana-6094	49	39	of	of	ADP
cana-6094	49	40	mathematical	mathematical	ADJ
cana-6094	49	41	objects	object	NOUN
cana-6094	49	42	.	.	PUNCT
cana-6094	49	43	3.2	3.2	NUM
cana-6094	49	44	spatial	spatial	ADJ
cana-6094	49	45	graph	graph	NOUN
cana-6094	49	46	convolutional	convolutional	ADJ
cana-6094	49	47	networks	network	NOUN
cana-6094	49	48	spatial	spatial	ADJ
cana-6094	49	49	gnns	gnn	NOUN
cana-6094	49	50	define	define	VERB
cana-6094	49	51	convolution	convolution	NOUN
cana-6094	49	52	operations	operation	NOUN
cana-6094	49	53	directly	directly	ADV
cana-6094	49	54	in	in	ADP
cana-6094	49	55	the	the	DET
cana-6094	49	56	vertex	vertex	NOUN
cana-6094	49	57	domain	domain	NOUN
cana-6094	49	58	by	by	ADP
cana-6094	49	59	aggregating	aggregate	VERB
cana-6094	49	60	information	information	NOUN
cana-6094	49	61	from	from	ADP
cana-6094	49	62	local	local	ADJ
cana-6094	49	63	neighborhoods	neighborhood	NOUN
cana-6094	49	64	.	.	PUNCT
cana-6094	50	1	unlike	unlike	ADP
cana-6094	50	2	spectral	spectral	ADJ
cana-6094	50	3	methods	method	NOUN
cana-6094	50	4	that	that	PRON
cana-6094	50	5	require	require	VERB
cana-6094	50	6	expensive	expensive	ADJ
cana-6094	50	7	spectral	spectral	ADJ
cana-6094	50	8	decomposition	decomposition	NOUN
cana-6094	50	9	,	,	PUNCT
cana-6094	50	10	spatial	spatial	ADJ
cana-6094	50	11	approaches	approach	NOUN
cana-6094	50	12	are	be	AUX
cana-6094	50	13	computationally	computationally	ADV
cana-6094	50	14	efficient	efficient	ADJ
cana-6094	50	15	and	and	CCONJ
cana-6094	50	16	can	can	AUX
cana-6094	50	17	be	be	AUX
cana-6094	50	18	applied	apply	VERB
cana-6094	50	19	to	to	ADP
cana-6094	50	20	large	large	ADJ
cana-6094	50	21	mathematical	mathematical	ADJ
cana-6094	50	22	structures	structure	NOUN
cana-6094	50	23	14	14	NUM
cana-6094	50	24	.	.	PUNCT
cana-6094	51	1	the	the	DET
cana-6094	51	2	graph	graph	NOUN
cana-6094	51	3	convolutional	convolutional	ADJ
cana-6094	51	4	network	network	NOUN
cana-6094	51	5	(	(	PUNCT
cana-6094	51	6	gcn	gcn	NOUN
cana-6094	51	7	)	)	PUNCT
cana-6094	51	8	is	be	AUX
cana-6094	51	9	a	a	DET
cana-6094	51	10	seminal	seminal	ADJ
cana-6094	51	11	spatial	spatial	ADJ
cana-6094	51	12	approach	approach	NOUN
cana-6094	51	13	that	that	PRON
cana-6094	51	14	simplifies	simplifie	NOUN
cana-6094	51	15	convolution	convolution	NOUN
cana-6094	51	16	as	as	ADP
cana-6094	51	17	a	a	DET
cana-6094	51	18	neighborhood	neighborhood	NOUN
cana-6094	51	19	aggregation	aggregation	NOUN
cana-6094	51	20	step	step	NOUN
cana-6094	51	21	:	:	PUNCT
cana-6094	51	22	h(l+1)=σ(d~−21a	h(l+1)=σ(d~−21a	ADJ
cana-6094	51	23	~	~	X
cana-6094	51	24	d~−21h(l)w(l	d~−21h(l)w(l	X
cana-6094	51	25	)	)	PUNCT
cana-6094	51	26	)	)	PUNCT
cana-6094	51	27	where	where	SCONJ
cana-6094	51	28	{	{	PUNCT
cana-6094	51	29	a	a	NOUN
cana-6094	51	30	}	}	PUNCT
cana-6094	51	31	=	=	PUNCT
cana-6094	51	32	a	a	PRON
cana-6094	52	1	+	+	CCONJ
cana-6094	52	2	i	i	PRON
cana-6094	52	3	is	be	AUX
cana-6094	52	4	the	the	DET
cana-6094	52	5	adjacency	adjacency	NOUN
cana-6094	52	6	matrix	matrix	NOUN
cana-6094	52	7	with	with	ADP
cana-6094	52	8	self	self	NOUN
cana-6094	52	9	-	-	PUNCT
cana-6094	52	10	loops	loop	NOUN
cana-6094	52	11	,	,	PUNCT
cana-6094	52	12	{	{	PUNCT
cana-6094	52	13	d	d	X
cana-6094	52	14	}	}	PUNCT
cana-6094	52	15	is	be	AUX
cana-6094	52	16	the	the	DET
cana-6094	52	17	corresponding	correspond	VERB
cana-6094	52	18	degree	degree	NOUN
cana-6094	52	19	matrix	matrix	NOUN
cana-6094	52	20	,	,	PUNCT
cana-6094	52	21	h^{(l	h^{(l	NOUN
cana-6094	52	22	)	)	PUNCT
cana-6094	52	23	}	}	PUNCT
cana-6094	52	24	is	be	AUX
cana-6094	52	25	the	the	DET
cana-6094	52	26	node	node	ADJ
cana-6094	52	27	representation	representation	NOUN
cana-6094	52	28	at	at	ADP
cana-6094	52	29	layer	layer	NOUN
cana-6094	52	30	l	l	NOUN
cana-6094	52	31	,	,	PUNCT
cana-6094	52	32	and	and	CCONJ
cana-6094	52	33	w^{(l	w^{(l	NOUN
cana-6094	52	34	)	)	PUNCT
cana-6094	52	35	}	}	PUNCT
cana-6094	52	36	is	be	AUX
cana-6094	52	37	a	a	DET
cana-6094	52	38	learnable	learnable	ADJ
cana-6094	52	39	weight	weight	NOUN
cana-6094	52	40	matrix	matrix	NOUN
cana-6094	52	41	1	1	NUM
cana-6094	52	42	.	.	PUNCT
cana-6094	53	1	this	this	DET
cana-6094	53	2	formulation	formulation	NOUN
cana-6094	53	3	represents	represent	VERB
cana-6094	53	4	a	a	DET
cana-6094	53	5	first	first	ADJ
cana-6094	53	6	-	-	PUNCT
cana-6094	53	7	order	order	NOUN
cana-6094	53	8	approximation	approximation	NOUN
cana-6094	53	9	of	of	ADP
cana-6094	53	10	spectral	spectral	ADJ
cana-6094	53	11	convolution	convolution	NOUN
cana-6094	53	12	while	while	SCONJ
cana-6094	53	13	being	be	AUX
cana-6094	53	14	computationally	computationally	ADV
cana-6094	53	15	efficient	efficient	ADJ
cana-6094	53	16	.	.	PUNCT
cana-6094	54	1	for	for	ADP
cana-6094	54	2	mathematical	mathematical	ADJ
cana-6094	54	3	analysis	analysis	NOUN
cana-6094	54	4	,	,	PUNCT
cana-6094	54	5	spatial	spatial	ADJ
cana-6094	54	6	gnns	gnns	NOUN
cana-6094	54	7	excel	excel	VERB
cana-6094	54	8	at	at	ADP
cana-6094	54	9	capturing	capture	VERB
cana-6094	54	10	local	local	ADJ
cana-6094	54	11	patterns	pattern	NOUN
cana-6094	54	12	and	and	CCONJ
cana-6094	54	13	propagation	propagation	NOUN
cana-6094	54	14	properties	property	NOUN
cana-6094	54	15	.	.	PUNCT
cana-6094	55	1	they	they	PRON
cana-6094	55	2	are	be	AUX
cana-6094	55	3	particularly	particularly	ADV
cana-6094	55	4	effective	effective	ADJ
cana-6094	55	5	for	for	ADP
cana-6094	55	6	structures	structure	NOUN
cana-6094	55	7	where	where	SCONJ
cana-6094	55	8	local	local	ADJ
cana-6094	55	9	relations	relation	NOUN
cana-6094	55	10	determine	determine	VERB
cana-6094	55	11	global	global	ADJ
cana-6094	55	12	properties	property	NOUN
cana-6094	55	13	,	,	PUNCT
cana-6094	55	14	such	such	ADJ
cana-6094	55	15	as	as	ADP
cana-6094	55	16	in	in	ADP
cana-6094	55	17	algebraic	algebraic	ADJ
cana-6094	55	18	structures	structure	NOUN
cana-6094	55	19	where	where	SCONJ
cana-6094	55	20	group	group	NOUN
cana-6094	55	21	properties	property	NOUN
cana-6094	55	22	emerge	emerge	VERB
cana-6094	55	23	from	from	ADP
cana-6094	55	24	generator	generator	NOUN
cana-6094	55	25	applications	application	NOUN
cana-6094	55	26	,	,	PUNCT
cana-6094	55	27	or	or	CCONJ
cana-6094	55	28	in	in	ADP
cana-6094	55	29	geometric	geometric	ADJ
cana-6094	55	30	structures	structure	NOUN
cana-6094	55	31	where	where	SCONJ
cana-6094	55	32	local	local	ADJ
cana-6094	55	33	metrics	metric	NOUN
cana-6094	55	34	determine	determine	VERB
cana-6094	55	35	global	global	ADJ
cana-6094	55	36	geometry	geometry	NOUN
cana-6094	55	37	47	47	NUM
cana-6094	55	38	.	.	PUNCT
cana-6094	55	39	3.3	3.3	NUM
cana-6094	55	40	attention	attention	NOUN
cana-6094	55	41	-	-	PUNCT
cana-6094	55	42	based	base	VERB
cana-6094	55	43	and	and	CCONJ
cana-6094	55	44	advanced	advanced	ADJ
cana-6094	55	45	architectures	architecture	NOUN
cana-6094	55	46	graph	graph	VERB
cana-6094	55	47	attention	attention	NOUN
cana-6094	55	48	networks	network	NOUN
cana-6094	55	49	(	(	PUNCT
cana-6094	55	50	gats	gat	NOUN
cana-6094	55	51	)	)	PUNCT
cana-6094	55	52	introduce	introduce	VERB
cana-6094	55	53	attention	attention	NOUN
cana-6094	55	54	mechanisms	mechanism	NOUN
cana-6094	55	55	to	to	ADP
cana-6094	55	56	gnns	gnns	NOUN
cana-6094	55	57	,	,	PUNCT
cana-6094	55	58	allowing	allow	VERB
cana-6094	55	59	models	model	NOUN
cana-6094	55	60	to	to	PART
cana-6094	55	61	learn	learn	VERB
cana-6094	55	62	adaptive	adaptive	ADJ
cana-6094	55	63	weights	weight	NOUN
cana-6094	55	64	for	for	ADP
cana-6094	55	65	neighborhood	neighborhood	NOUN
cana-6094	55	66	aggregation	aggregation	NOUN
cana-6094	55	67	.	.	PUNCT
cana-6094	56	1	for	for	ADP
cana-6094	56	2	each	each	DET
cana-6094	56	3	edge	edge	NOUN
cana-6094	56	4	(	(	PUNCT
cana-6094	56	5	i	i	PROPN
cana-6094	56	6	,	,	PUNCT
cana-6094	56	7	j	j	PROPN
cana-6094	56	8	)	)	PUNCT
cana-6094	56	9	,	,	PUNCT
cana-6094	56	10	gat	gat	NOUN
cana-6094	56	11	computes	compute	VERB
cana-6094	56	12	an	an	DET
cana-6094	56	13	attention	attention	NOUN
cana-6094	56	14	coefficient	coefficient	NOUN
cana-6094	56	15	:	:	PUNCT
cana-6094	56	16	αij=∑k∈n(i)exp(leakyrelu(at[whi∣∣whk]))exp(leakyrelu(at[whi∣∣whj	αij=∑k∈n(i)exp(leakyrelu(at[whi∣∣whk]))exp(leakyrelu(at[whi∣∣whj	X
cana-6094	56	17	]	]	X
cana-6094	56	18	)	)	PUNCT
cana-6094	56	19	)	)	PUNCT
cana-6094	56	20	where	where	SCONJ
cana-6094	56	21	a	a	PRON
cana-6094	56	22	and	and	CCONJ
cana-6094	56	23	w	w	NOUN
cana-6094	56	24	are	be	AUX
cana-6094	56	25	learnable	learnable	ADJ
cana-6094	56	26	parameters	parameter	NOUN
cana-6094	56	27	,	,	PUNCT
cana-6094	56	28	and	and	CCONJ
cana-6094	56	29	||	||	NUM
cana-6094	56	30	denotes	denote	NOUN
cana-6094	56	31	concatenation	concatenation	VERB
cana-6094	56	32	3	3	NUM
cana-6094	56	33	.	.	PUNCT
cana-6094	57	1	the	the	DET
cana-6094	57	2	node	node	ADJ
cana-6094	57	3	update	update	NOUN
cana-6094	57	4	then	then	ADV
cana-6094	57	5	becomes	become	VERB
cana-6094	57	6	a	a	DET
cana-6094	57	7	weighted	weighted	ADJ
cana-6094	57	8	aggregation	aggregation	NOUN
cana-6094	57	9	of	of	ADP
cana-6094	57	10	neighborhood	neighborhood	NOUN
cana-6094	57	11	features	feature	NOUN
cana-6094	57	12	based	base	VERB
cana-6094	57	13	on	on	ADP
cana-6094	57	14	these	these	DET
cana-6094	57	15	attention	attention	NOUN
cana-6094	57	16	weights	weight	NOUN
cana-6094	57	17	.	.	PUNCT
cana-6094	58	1	for	for	ADP
cana-6094	58	2	mathematical	mathematical	ADJ
cana-6094	58	3	structures	structure	NOUN
cana-6094	58	4	,	,	PUNCT
cana-6094	58	5	attention	attention	NOUN
cana-6094	58	6	mechanisms	mechanism	NOUN
cana-6094	58	7	provide	provide	VERB
cana-6094	58	8	several	several	ADJ
cana-6094	58	9	advantages	advantage	NOUN
cana-6094	58	10	:	:	PUNCT
cana-6094	58	11	•	•	SCONJ
cana-6094	58	12	they	they	PRON
cana-6094	58	13	can	can	AUX
cana-6094	58	14	learn	learn	VERB
cana-6094	58	15	to	to	PART
cana-6094	58	16	focus	focus	VERB
cana-6094	58	17	on	on	ADP
cana-6094	58	18	structurally	structurally	ADV
cana-6094	58	19	important	important	ADJ
cana-6094	58	20	relationships	relationship	NOUN
cana-6094	58	21	,	,	PUNCT
cana-6094	58	22	such	such	ADJ
cana-6094	58	23	as	as	ADP
cana-6094	58	24	key	key	ADJ
cana-6094	58	25	generators	generator	NOUN
cana-6094	58	26	in	in	ADP
cana-6094	58	27	algebraic	algebraic	ADJ
cana-6094	58	28	structures	structure	NOUN
cana-6094	58	29	or	or	CCONJ
cana-6094	58	30	critical	critical	ADJ
cana-6094	58	31	connections	connection	NOUN
cana-6094	58	32	in	in	ADP
cana-6094	58	33	topological	topological	ADJ
cana-6094	58	34	spaces	space	NOUN
cana-6094	58	35	.	.	PUNCT
cana-6094	59	1	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	PROPN
cana-6094	59	2	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	PROPN
cana-6094	59	3	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	https://pmc.ncbi.nlm.nih.gov/articles/pmc8360394/	PROPN
cana-6094	59	4	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	PROPN
cana-6094	59	5	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	VERB
cana-6094	59	6	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	59	7	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	NOUN
cana-6094	59	8	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	X
cana-6094	59	9	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	X
cana-6094	59	10	communications	communication	NOUN
cana-6094	59	11	on	on	ADP
cana-6094	59	12	applied	apply	VERB
cana-6094	59	13	nonlinear	nonlinear	ADJ
cana-6094	59	14	analysis	analysis	NOUN
cana-6094	59	15	issn	issn	NOUN
cana-6094	59	16	:	:	PUNCT
cana-6094	59	17	1074	1074	NUM
cana-6094	59	18	-	-	PUNCT
cana-6094	59	19	133x	133x	NUM
cana-6094	59	20	vol	vol	NOUN
cana-6094	59	21	29	29	NUM
cana-6094	59	22	no	no	NOUN
cana-6094	59	23	.	.	NOUN
cana-6094	59	24	4	4	NUM
cana-6094	59	25	(	(	PUNCT
cana-6094	59	26	2022	2022	NUM
cana-6094	59	27	)	)	PUNCT
cana-6094	59	28	31	31	NUM
cana-6094	59	29	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	59	30	•	•	NOUN
cana-6094	59	31	they	they	PRON
cana-6094	59	32	enable	enable	VERB
cana-6094	59	33	interpretable	interpretable	ADJ
cana-6094	59	34	analysis	analysis	NOUN
cana-6094	59	35	by	by	ADP
cana-6094	59	36	revealing	reveal	VERB
cana-6094	59	37	which	which	PRON
cana-6094	59	38	relationships	relationship	NOUN
cana-6094	59	39	contribute	contribute	VERB
cana-6094	59	40	most	most	ADV
cana-6094	59	41	to	to	ADP
cana-6094	59	42	mathematical	mathematical	ADJ
cana-6094	59	43	properties	property	NOUN
cana-6094	59	44	.	.	PUNCT
cana-6094	60	1	•	•	NOUN
cana-6094	60	2	they	they	PRON
cana-6094	60	3	adapt	adapt	VERB
cana-6094	60	4	naturally	naturally	ADV
cana-6094	60	5	to	to	ADP
cana-6094	60	6	heterogeneous	heterogeneous	ADJ
cana-6094	60	7	mathematical	mathematical	ADJ
cana-6094	60	8	structures	structure	NOUN
cana-6094	60	9	where	where	SCONJ
cana-6094	60	10	different	different	ADJ
cana-6094	60	11	relationships	relationship	NOUN
cana-6094	60	12	have	have	VERB
cana-6094	60	13	varying	vary	VERB
cana-6094	60	14	importance	importance	NOUN
cana-6094	60	15	3	3	NUM
cana-6094	60	16	.	.	PUNCT
cana-6094	60	17	advanced	advanced	ADJ
cana-6094	60	18	architectures	architecture	NOUN
cana-6094	60	19	like	like	ADP
cana-6094	60	20	graph	graph	NOUN
cana-6094	60	21	transformers	transformer	NOUN
cana-6094	60	22	extend	extend	VERB
cana-6094	60	23	attention	attention	NOUN
cana-6094	60	24	mechanisms	mechanism	NOUN
cana-6094	60	25	to	to	PART
cana-6094	60	26	capture	capture	VERB
cana-6094	60	27	global	global	ADJ
cana-6094	60	28	dependencies	dependency	NOUN
cana-6094	60	29	beyond	beyond	ADP
cana-6094	60	30	local	local	ADJ
cana-6094	60	31	neighborhoods	neighborhood	NOUN
cana-6094	60	32	,	,	PUNCT
cana-6094	60	33	which	which	PRON
cana-6094	60	34	is	be	AUX
cana-6094	60	35	particularly	particularly	ADV
cana-6094	60	36	valuable	valuable	ADJ
cana-6094	60	37	for	for	ADP
cana-6094	60	38	mathematical	mathematical	ADJ
cana-6094	60	39	structures	structure	NOUN
cana-6094	60	40	with	with	ADP
cana-6094	60	41	long	long	ADJ
cana-6094	60	42	-	-	PUNCT
cana-6094	60	43	range	range	NOUN
cana-6094	60	44	dependencies	dependency	NOUN
cana-6094	60	45	or	or	CCONJ
cana-6094	60	46	global	global	ADJ
cana-6094	60	47	symmetries	symmetry	NOUN
cana-6094	60	48	38	38	NUM
cana-6094	60	49	.	.	PUNCT
cana-6094	61	1	3.4	3.4	NUM
cana-6094	61	2	geometrically	geometrically	ADV
cana-6094	61	3	equivariant	equivariant	ADJ
cana-6094	61	4	gnns	gnns	PROPN
cana-6094	61	5	geometrically	geometrically	ADV
cana-6094	61	6	equivariant	equivariant	ADJ
cana-6094	61	7	gnns	gnns	PROPN
cana-6094	61	8	specialize	specialize	NOUN
cana-6094	61	9	in	in	ADP
cana-6094	61	10	processing	process	VERB
cana-6094	61	11	geometric	geometric	ADJ
cana-6094	61	12	graphs	graph	NOUN
cana-6094	61	13	while	while	SCONJ
cana-6094	61	14	preserving	preserve	VERB
cana-6094	61	15	important	important	ADJ
cana-6094	61	16	mathematical	mathematical	ADJ
cana-6094	61	17	symmetries	symmetry	NOUN
cana-6094	61	18	such	such	ADJ
cana-6094	61	19	as	as	ADP
cana-6094	61	20	translation	translation	NOUN
cana-6094	61	21	,	,	PUNCT
cana-6094	61	22	rotation	rotation	NOUN
cana-6094	61	23	,	,	PUNCT
cana-6094	61	24	and	and	CCONJ
cana-6094	61	25	reflection	reflection	NOUN
cana-6094	61	26	invariance	invariance	NOUN
cana-6094	61	27	.	.	PUNCT
cana-6094	62	1	these	these	DET
cana-6094	62	2	architectures	architecture	NOUN
cana-6094	62	3	are	be	AUX
cana-6094	62	4	particularly	particularly	ADV
cana-6094	62	5	valuable	valuable	ADJ
cana-6094	62	6	for	for	ADP
cana-6094	62	7	analyzing	analyze	VERB
cana-6094	62	8	geometric	geometric	ADJ
cana-6094	62	9	and	and	CCONJ
cana-6094	62	10	topological	topological	ADJ
cana-6094	62	11	structures	structure	NOUN
cana-6094	62	12	where	where	SCONJ
cana-6094	62	13	preserving	preserve	VERB
cana-6094	62	14	geometric	geometric	ADJ
cana-6094	62	15	relationships	relationship	NOUN
cana-6094	62	16	is	be	AUX
cana-6094	62	17	essential	essential	ADJ
cana-6094	62	18	8	8	NUM
cana-6094	62	19	.	.	PUNCT
cana-6094	63	1	the	the	DET
cana-6094	63	2	core	core	ADJ
cana-6094	63	3	idea	idea	NOUN
cana-6094	63	4	behind	behind	ADP
cana-6094	63	5	equivariant	equivariant	PROPN
cana-6094	63	6	gnns	gnns	PROPN
cana-6094	63	7	is	be	AUX
cana-6094	63	8	to	to	PART
cana-6094	63	9	ensure	ensure	VERB
cana-6094	63	10	that	that	SCONJ
cana-6094	63	11	network	network	NOUN
cana-6094	63	12	transformations	transformation	NOUN
cana-6094	63	13	commute	commute	VERB
cana-6094	63	14	with	with	ADP
cana-6094	63	15	symmetry	symmetry	NOUN
cana-6094	63	16	operations	operation	NOUN
cana-6094	63	17	:	:	PUNCT
cana-6094	63	18	f(ρ(g)⋅x)=ρ′(g)⋅f(x	f(ρ(g)⋅x)=ρ′(g)⋅f(x	NOUN
cana-6094	63	19	)	)	PUNCT
cana-6094	63	20	where	where	SCONJ
cana-6094	63	21	f	f	PROPN
cana-6094	63	22	is	be	AUX
cana-6094	63	23	the	the	DET
cana-6094	63	24	network	network	NOUN
cana-6094	63	25	transformation	transformation	NOUN
cana-6094	63	26	are	be	AUX
cana-6094	63	27	representations	representation	NOUN
cana-6094	63	28	of	of	ADP
cana-6094	63	29	the	the	DET
cana-6094	63	30	symmetry	symmetry	NOUN
cana-6094	63	31	group	group	NOUN
cana-6094	63	32	g	g	PROPN
cana-6094	63	33	,	,	PUNCT
cana-6094	63	34	and	and	CCONJ
cana-6094	63	35	x	x	X
cana-6094	63	36	is	be	AUX
cana-6094	63	37	the	the	DET
cana-6094	63	38	input	input	NOUN
cana-6094	63	39	8	8	NUM
cana-6094	63	40	.	.	PUNCT
cana-6094	64	1	this	this	DET
cana-6094	64	2	property	property	NOUN
cana-6094	64	3	ensures	ensure	VERB
cana-6094	64	4	that	that	SCONJ
cana-6094	64	5	the	the	DET
cana-6094	64	6	analysis	analysis	NOUN
cana-6094	64	7	respects	respect	VERB
cana-6094	64	8	the	the	DET
cana-6094	64	9	inherent	inherent	ADJ
cana-6094	64	10	symmetries	symmetry	NOUN
cana-6094	64	11	of	of	ADP
cana-6094	64	12	mathematical	mathematical	ADJ
cana-6094	64	13	structures	structure	NOUN
cana-6094	64	14	.	.	PUNCT
cana-6094	65	1	e(n)-equivariant	e(n)-equivariant	ADJ
cana-6094	65	2	gnns	gnns	NOUN
cana-6094	65	3	specifically	specifically	ADV
cana-6094	65	4	preserve	preserve	VERB
cana-6094	65	5	equivariance	equivariance	NOUN
cana-6094	65	6	under	under	ADP
cana-6094	65	7	the	the	DET
cana-6094	65	8	euclidean	euclidean	ADJ
cana-6094	65	9	group	group	NOUN
cana-6094	65	10	e(n	e(n	PROPN
cana-6094	65	11	)	)	PUNCT
cana-6094	65	12	of	of	ADP
cana-6094	65	13	rotations	rotation	NOUN
cana-6094	65	14	,	,	PUNCT
cana-6094	65	15	translations	translation	NOUN
cana-6094	65	16	,	,	PUNCT
cana-6094	65	17	and	and	CCONJ
cana-6094	65	18	reflections	reflection	NOUN
cana-6094	65	19	in	in	ADP
cana-6094	65	20	n	n	ADV
cana-6094	65	21	-	-	PUNCT
cana-6094	65	22	dimensional	dimensional	ADJ
cana-6094	65	23	space	space	NOUN
cana-6094	65	24	.	.	PUNCT
cana-6094	66	1	these	these	DET
cana-6094	66	2	networks	network	NOUN
cana-6094	66	3	have	have	AUX
cana-6094	66	4	proven	prove	VERB
cana-6094	66	5	particularly	particularly	ADV
cana-6094	66	6	effective	effective	ADJ
cana-6094	66	7	for	for	ADP
cana-6094	66	8	analyzing	analyze	VERB
cana-6094	66	9	molecular	molecular	ADJ
cana-6094	66	10	structures	structure	NOUN
cana-6094	66	11	,	,	PUNCT
cana-6094	66	12	crystalline	crystalline	ADJ
cana-6094	66	13	materials	material	NOUN
cana-6094	66	14	,	,	PUNCT
cana-6094	66	15	and	and	CCONJ
cana-6094	66	16	geometric	geometric	ADJ
cana-6094	66	17	shapes	shape	NOUN
cana-6094	66	18	—	—	PUNCT
cana-6094	66	19	all	all	PRON
cana-6094	66	20	of	of	ADP
cana-6094	66	21	which	which	PRON
cana-6094	66	22	are	be	AUX
cana-6094	66	23	mathematical	mathematical	ADJ
cana-6094	66	24	structures	structure	NOUN
cana-6094	66	25	with	with	ADP
cana-6094	66	26	rich	rich	ADJ
cana-6094	66	27	geometric	geometric	ADJ
cana-6094	66	28	properties	property	NOUN
cana-6094	66	29	8	8	NUM
cana-6094	66	30	.	.	PUNCT
cana-6094	66	31	architecture	architecture	NOUN
cana-6094	66	32	type	type	NOUN
cana-6094	66	33	key	key	ADJ
cana-6094	66	34	mechanism	mechanism	NOUN
cana-6094	66	35	mathematical	mathematical	ADJ
cana-6094	66	36	strengths	strength	NOUN
cana-6094	66	37	application	application	NOUN
cana-6094	66	38	examples	example	NOUN
cana-6094	67	1	spectral	spectral	ADJ
cana-6094	67	2	gnns	gnns	PROPN
cana-6094	67	3	graph	graph	NOUN
cana-6094	67	4	fourier	fourier	NOUN
cana-6094	67	5	transform	transform	VERB
cana-6094	67	6	global	global	ADJ
cana-6094	67	7	symmetry	symmetry	NOUN
cana-6094	67	8	analysis	analysis	NOUN
cana-6094	67	9	,	,	PUNCT
cana-6094	67	10	invariant	invariant	ADJ
cana-6094	67	11	properties	property	NOUN
cana-6094	67	12	isospectral	isospectral	ADJ
cana-6094	67	13	graph	graph	NOUN
cana-6094	67	14	analysis	analysis	NOUN
cana-6094	67	15	,	,	PUNCT
cana-6094	67	16	symmetry	symmetry	NOUN
cana-6094	67	17	detection	detection	NOUN
cana-6094	67	18	spatial	spatial	ADJ
cana-6094	67	19	gnns	gnns	ADJ
cana-6094	67	20	neighborhood	neighborhood	NOUN
cana-6094	67	21	aggregation	aggregation	NOUN
cana-6094	67	22	local	local	ADJ
cana-6094	67	23	pattern	pattern	NOUN
cana-6094	67	24	recognition	recognition	NOUN
cana-6094	67	25	,	,	PUNCT
cana-6094	67	26	propagation	propagation	NOUN
cana-6094	67	27	properties	property	NOUN
cana-6094	67	28	group	group	NOUN
cana-6094	67	29	property	property	NOUN
cana-6094	67	30	prediction	prediction	NOUN
cana-6094	67	31	,	,	PUNCT
cana-6094	67	32	combinatorial	combinatorial	ADJ
cana-6094	67	33	optimization	optimization	NOUN
cana-6094	67	34	attention	attention	NOUN
cana-6094	67	35	-	-	PUNCT
cana-6094	67	36	based	base	VERB
cana-6094	67	37	gnns	gnns	ADJ
cana-6094	67	38	adaptive	adaptive	ADJ
cana-6094	67	39	neighborhood	neighborhood	NOUN
cana-6094	67	40	weighting	weight	VERB
cana-6094	67	41	importance	importance	NOUN
cana-6094	67	42	learning	learning	NOUN
cana-6094	67	43	,	,	PUNCT
cana-6094	67	44	interpretable	interpretable	ADJ
cana-6094	67	45	analysis	analysis	NOUN
cana-6094	67	46	key	key	ADJ
cana-6094	67	47	generator	generator	NOUN
cana-6094	67	48	identification	identification	NOUN
cana-6094	67	49	,	,	PUNCT
cana-6094	67	50	structure	structure	NOUN
cana-6094	67	51	interpretation	interpretation	NOUN
cana-6094	67	52	equivariant	equivariant	PROPN
cana-6094	67	53	gnns	gnns	PROPN
cana-6094	67	54	symmetry	symmetry	NOUN
cana-6094	67	55	-	-	PUNCT
cana-6094	67	56	preserving	preserve	VERB
cana-6094	67	57	transformations	transformation	NOUN
cana-6094	67	58	geometric	geometric	ADJ
cana-6094	67	59	property	property	NOUN
cana-6094	67	60	preservation	preservation	NOUN
cana-6094	67	61	,	,	PUNCT
cana-6094	67	62	coordinate	coordinate	VERB
cana-6094	67	63	independence	independence	NOUN
cana-6094	67	64	molecular	molecular	ADJ
cana-6094	67	65	geometry	geometry	NOUN
cana-6094	67	66	,	,	PUNCT
cana-6094	67	67	crystal	crystal	NOUN
cana-6094	67	68	structure	structure	NOUN
cana-6094	67	69	analysis	analysis	NOUN
cana-6094	67	70	petri	petri	ADJ
cana-6094	67	71	net	net	ADJ
cana-6094	67	72	gnns	gnns	ADJ
cana-6094	67	73	multimodal	multimodal	NOUN
cana-6094	67	74	network	network	NOUN
cana-6094	67	75	flows	flow	VERB
cana-6094	67	76	higher	high	ADJ
cana-6094	67	77	-	-	PUNCT
cana-6094	67	78	order	order	NOUN
cana-6094	67	79	interactions	interaction	NOUN
cana-6094	67	80	,	,	PUNCT
cana-6094	67	81	concurrent	concurrent	ADJ
cana-6094	67	82	processes	process	NOUN
cana-6094	67	83	complex	complex	ADJ
cana-6094	67	84	system	system	NOUN
cana-6094	67	85	modeling	modeling	NOUN
cana-6094	67	86	,	,	PUNCT
cana-6094	67	87	chemical	chemical	ADJ
cana-6094	67	88	reactions	reaction	NOUN
cana-6094	67	89	table	table	VERB
cana-6094	67	90	2	2	NUM
cana-6094	67	91	:	:	PUNCT
cana-6094	67	92	gnn	gnn	PROPN
cana-6094	67	93	architectures	architecture	NOUN
cana-6094	67	94	for	for	ADP
cana-6094	67	95	mathematical	mathematical	ADJ
cana-6094	67	96	analysis	analysis	NOUN
cana-6094	67	97	4.applications	4.application	NOUN
cana-6094	67	98	to	to	ADP
cana-6094	67	99	mathematical	mathematical	ADJ
cana-6094	67	100	structures	structure	NOUN
cana-6094	67	101	4.1	4.1	NUM
cana-6094	67	102	analysis	analysis	NOUN
cana-6094	67	103	of	of	ADP
cana-6094	67	104	algebraic	algebraic	ADJ
cana-6094	67	105	structures	structure	NOUN
cana-6094	67	106	gnns	gnn	NOUN
cana-6094	67	107	have	have	AUX
cana-6094	67	108	demonstrated	demonstrate	VERB
cana-6094	67	109	remarkable	remarkable	ADJ
cana-6094	67	110	capabilities	capability	NOUN
cana-6094	67	111	in	in	ADP
cana-6094	67	112	analyzing	analyze	VERB
cana-6094	67	113	algebraic	algebraic	ADJ
cana-6094	67	114	structures	structure	NOUN
cana-6094	67	115	such	such	ADJ
cana-6094	67	116	as	as	ADP
cana-6094	67	117	groups	group	NOUN
cana-6094	67	118	,	,	PUNCT
cana-6094	67	119	rings	ring	NOUN
cana-6094	67	120	,	,	PUNCT
cana-6094	67	121	and	and	CCONJ
cana-6094	67	122	fields	field	NOUN
cana-6094	67	123	by	by	ADP
cana-6094	67	124	learning	learn	VERB
cana-6094	67	125	from	from	ADP
cana-6094	67	126	their	their	PRON
cana-6094	67	127	graph	graph	NOUN
cana-6094	67	128	representations	representation	NOUN
cana-6094	67	129	.	.	PUNCT
cana-6094	68	1	for	for	ADP
cana-6094	68	2	group	group	NOUN
cana-6094	68	3	theory	theory	NOUN
cana-6094	68	4	,	,	PUNCT
cana-6094	68	5	cayley	cayley	ADJ
cana-6094	68	6	graphs	graph	NOUN
cana-6094	68	7	provide	provide	VERB
cana-6094	68	8	a	a	DET
cana-6094	68	9	natural	natural	ADJ
cana-6094	68	10	representation	representation	NOUN
cana-6094	68	11	where	where	SCONJ
cana-6094	68	12	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	PROPN
cana-6094	68	13	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	PROPN
cana-6094	69	1	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	69	2	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	69	3	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	69	4	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	69	5	communications	communication	NOUN
cana-6094	69	6	on	on	ADP
cana-6094	69	7	applied	apply	VERB
cana-6094	69	8	nonlinear	nonlinear	ADJ
cana-6094	69	9	analysis	analysis	NOUN
cana-6094	69	10	issn	issn	NOUN
cana-6094	69	11	:	:	PUNCT
cana-6094	69	12	1074	1074	NUM
cana-6094	69	13	-	-	PUNCT
cana-6094	69	14	133x	133x	NUM
cana-6094	69	15	vol	vol	NOUN
cana-6094	69	16	29	29	NUM
cana-6094	69	17	no	no	NOUN
cana-6094	69	18	.	.	NOUN
cana-6094	69	19	4	4	NUM
cana-6094	69	20	(	(	PUNCT
cana-6094	69	21	2022	2022	NUM
cana-6094	69	22	)	)	PUNCT
cana-6094	69	23	32	32	NUM
cana-6094	69	24	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	69	25	nodes	node	NOUN
cana-6094	69	26	correspond	correspond	VERB
cana-6094	69	27	to	to	ADP
cana-6094	69	28	group	group	NOUN
cana-6094	69	29	elements	element	NOUN
cana-6094	69	30	and	and	CCONJ
cana-6094	69	31	edges	edge	NOUN
cana-6094	69	32	represent	represent	VERB
cana-6094	69	33	generator	generator	NOUN
cana-6094	69	34	applications	application	NOUN
cana-6094	69	35	.	.	PUNCT
cana-6094	70	1	gnns	gnns	PROPN
cana-6094	70	2	can	can	AUX
cana-6094	70	3	learn	learn	VERB
cana-6094	70	4	to	to	PART
cana-6094	70	5	predict	predict	VERB
cana-6094	70	6	group	group	NOUN
cana-6094	70	7	properties	property	NOUN
cana-6094	70	8	such	such	ADJ
cana-6094	70	9	as	as	ADP
cana-6094	70	10	commutativity	commutativity	NOUN
cana-6094	70	11	,	,	PUNCT
cana-6094	70	12	solvability	solvability	NOUN
cana-6094	70	13	,	,	PUNCT
cana-6094	70	14	and	and	CCONJ
cana-6094	70	15	nilpotency	nilpotency	NOUN
cana-6094	70	16	directly	directly	ADV
cana-6094	70	17	from	from	ADP
cana-6094	70	18	these	these	DET
cana-6094	70	19	graph	graph	NOUN
cana-6094	70	20	representations	representation	NOUN
cana-6094	70	21	7	7	NUM
cana-6094	70	22	.	.	PUNCT
cana-6094	71	1	recent	recent	ADJ
cana-6094	71	2	work	work	NOUN
cana-6094	71	3	has	have	AUX
cana-6094	71	4	applied	apply	VERB
cana-6094	71	5	gnns	gnns	NOUN
cana-6094	71	6	to	to	ADP
cana-6094	71	7	automatic	automatic	ADJ
cana-6094	71	8	theorem	theorem	NOUN
cana-6094	71	9	proving	proving	NOUN
cana-6094	71	10	in	in	ADP
cana-6094	71	11	group	group	NOUN
cana-6094	71	12	theory	theory	NOUN
cana-6094	71	13	,	,	PUNCT
cana-6094	71	14	where	where	SCONJ
cana-6094	71	15	networks	network	NOUN
cana-6094	71	16	learn	learn	VERB
cana-6094	71	17	to	to	PART
cana-6094	71	18	predict	predict	VERB
cana-6094	71	19	the	the	DET
cana-6094	71	20	validity	validity	NOUN
cana-6094	71	21	of	of	ADP
cana-6094	71	22	group	group	NOUN
cana-6094	71	23	-	-	PUNCT
cana-6094	71	24	theoretic	theoretic	NOUN
cana-6094	71	25	statements	statement	NOUN
cana-6094	71	26	based	base	VERB
cana-6094	71	27	on	on	ADP
cana-6094	71	28	structural	structural	ADJ
cana-6094	71	29	patterns	pattern	NOUN
cana-6094	71	30	.	.	PUNCT
cana-6094	72	1	by	by	ADP
cana-6094	72	2	training	train	VERB
cana-6094	72	3	on	on	ADP
cana-6094	72	4	known	know	VERB
cana-6094	72	5	theorems	theorem	NOUN
cana-6094	72	6	and	and	CCONJ
cana-6094	72	7	their	their	PRON
cana-6094	72	8	proofs	proof	NOUN
cana-6094	72	9	,	,	PUNCT
cana-6094	72	10	gnns	gnns	PROPN
cana-6094	72	11	can	can	AUX
cana-6094	72	12	suggest	suggest	VERB
cana-6094	72	13	potential	potential	ADJ
cana-6094	72	14	proof	proof	ADJ
cana-6094	72	15	strategies	strategy	NOUN
cana-6094	72	16	or	or	CCONJ
cana-6094	72	17	identify	identify	VERB
cana-6094	72	18	counterexamples	counterexample	NOUN
cana-6094	72	19	to	to	PART
cana-6094	72	20	conjectures	conjecture	VERB
cana-6094	72	21	9	9	NUM
cana-6094	72	22	.	.	PUNCT
cana-6094	73	1	this	this	DET
cana-6094	73	2	approach	approach	NOUN
cana-6094	73	3	complements	complement	VERB
cana-6094	73	4	traditional	traditional	ADJ
cana-6094	73	5	symbolic	symbolic	ADJ
cana-6094	73	6	methods	method	NOUN
cana-6094	73	7	by	by	ADP
cana-6094	73	8	providing	provide	VERB
cana-6094	73	9	statistical	statistical	ADJ
cana-6094	73	10	evidence	evidence	NOUN
cana-6094	73	11	for	for	ADP
cana-6094	73	12	mathematical	mathematical	ADJ
cana-6094	73	13	claims	claim	NOUN
cana-6094	73	14	.	.	PUNCT
cana-6094	74	1	in	in	ADP
cana-6094	74	2	ring	ring	NOUN
cana-6094	74	3	and	and	CCONJ
cana-6094	74	4	field	field	NOUN
cana-6094	74	5	theory	theory	NOUN
cana-6094	74	6	,	,	PUNCT
cana-6094	74	7	gnns	gnn	NOUN
cana-6094	74	8	have	have	AUX
cana-6094	74	9	been	be	AUX
cana-6094	74	10	used	use	VERB
cana-6094	74	11	to	to	PART
cana-6094	74	12	analyze	analyze	VERB
cana-6094	74	13	algebraic	algebraic	ADJ
cana-6094	74	14	invariants	invariant	NOUN
cana-6094	74	15	and	and	CCONJ
cana-6094	74	16	properties	property	NOUN
cana-6094	74	17	.	.	PUNCT
cana-6094	75	1	for	for	ADP
cana-6094	75	2	example	example	NOUN
cana-6094	75	3	,	,	PUNCT
cana-6094	75	4	gnns	gnns	PROPN
cana-6094	75	5	can	can	AUX
cana-6094	75	6	predict	predict	VERB
cana-6094	75	7	whether	whether	SCONJ
cana-6094	75	8	a	a	DET
cana-6094	75	9	ring	ring	NOUN
cana-6094	75	10	is	be	AUX
cana-6094	75	11	an	an	DET
cana-6094	75	12	integral	integral	ADJ
cana-6094	75	13	domain	domain	NOUN
cana-6094	75	14	,	,	PUNCT
cana-6094	75	15	a	a	DET
cana-6094	75	16	principal	principal	ADJ
cana-6094	75	17	ideal	ideal	ADJ
cana-6094	75	18	domain	domain	NOUN
cana-6094	75	19	,	,	PUNCT
cana-6094	75	20	or	or	CCONJ
cana-6094	75	21	a	a	DET
cana-6094	75	22	unique	unique	ADJ
cana-6094	75	23	factorization	factorization	NOUN
cana-6094	75	24	domain	domain	NOUN
cana-6094	75	25	based	base	VERB
cana-6094	75	26	on	on	ADP
cana-6094	75	27	its	its	PRON
cana-6094	75	28	multiplicative	multiplicative	ADJ
cana-6094	75	29	structure	structure	NOUN
cana-6094	75	30	represented	represent	VERB
cana-6094	75	31	as	as	ADP
cana-6094	75	32	a	a	DET
cana-6094	75	33	graph	graph	NOUN
cana-6094	75	34	7	7	NUM
cana-6094	75	35	.	.	PUNCT
cana-6094	76	1	these	these	DET
cana-6094	76	2	applications	application	NOUN
cana-6094	76	3	demonstrate	demonstrate	VERB
cana-6094	76	4	how	how	SCONJ
cana-6094	76	5	gnns	gnns	NOUN
cana-6094	76	6	can	can	AUX
cana-6094	76	7	capture	capture	VERB
cana-6094	76	8	abstract	abstract	ADJ
cana-6094	76	9	algebraic	algebraic	ADJ
cana-6094	76	10	properties	property	NOUN
cana-6094	76	11	through	through	ADP
cana-6094	76	12	structural	structural	ADJ
cana-6094	76	13	pattern	pattern	NOUN
cana-6094	76	14	recognition	recognition	NOUN
cana-6094	76	15	.	.	PUNCT
cana-6094	77	1	4.2	4.2	NUM
cana-6094	77	2	geometric	geometric	ADJ
cana-6094	77	3	and	and	CCONJ
cana-6094	77	4	topological	topological	ADJ
cana-6094	77	5	analysis	analysis	NOUN
cana-6094	77	6	geometric	geometric	ADJ
cana-6094	77	7	and	and	CCONJ
cana-6094	77	8	topological	topological	ADJ
cana-6094	77	9	structures	structure	NOUN
cana-6094	77	10	present	present	VERB
cana-6094	77	11	unique	unique	ADJ
cana-6094	77	12	challenges	challenge	NOUN
cana-6094	77	13	for	for	ADP
cana-6094	77	14	analysis	analysis	NOUN
cana-6094	77	15	due	due	ADP
cana-6094	77	16	to	to	ADP
cana-6094	77	17	their	their	PRON
cana-6094	77	18	continuous	continuous	ADJ
cana-6094	77	19	nature	nature	NOUN
cana-6094	77	20	and	and	CCONJ
cana-6094	77	21	invariance	invariance	NOUN
cana-6094	77	22	requirements	requirement	NOUN
cana-6094	77	23	.	.	PUNCT
cana-6094	78	1	gnns	gnns	PROPN
cana-6094	78	2	have	have	AUX
cana-6094	78	3	been	be	AUX
cana-6094	78	4	successfully	successfully	ADV
cana-6094	78	5	applied	apply	VERB
cana-6094	78	6	to	to	ADP
cana-6094	78	7	discretized	discretized	ADJ
cana-6094	78	8	representations	representation	NOUN
cana-6094	78	9	of	of	ADP
cana-6094	78	10	these	these	DET
cana-6094	78	11	structures	structure	NOUN
cana-6094	78	12	,	,	PUNCT
cana-6094	78	13	enabling	enable	VERB
cana-6094	78	14	data	data	NOUN
cana-6094	78	15	-	-	PUNCT
cana-6094	78	16	driven	drive	VERB
cana-6094	78	17	geometric	geometric	ADJ
cana-6094	78	18	analysis	analysis	NOUN
cana-6094	78	19	8	8	NUM
cana-6094	78	20	.	.	PUNCT
cana-6094	79	1	for	for	ADP
cana-6094	79	2	topological	topological	ADJ
cana-6094	79	3	data	datum	NOUN
cana-6094	79	4	analysis	analysis	NOUN
cana-6094	79	5	,	,	PUNCT
cana-6094	79	6	gnns	gnns	ADJ
cana-6094	79	7	operate	operate	VERB
cana-6094	79	8	on	on	ADP
cana-6094	79	9	simplicial	simplicial	ADJ
cana-6094	79	10	complexes	complex	NOUN
cana-6094	79	11	and	and	CCONJ
cana-6094	79	12	cell	cell	NOUN
cana-6094	79	13	complexes	complex	NOUN
cana-6094	79	14	to	to	PART
cana-6094	79	15	compute	compute	VERB
cana-6094	79	16	topological	topological	ADJ
cana-6094	79	17	invariants	invariant	NOUN
cana-6094	79	18	such	such	ADJ
cana-6094	79	19	as	as	ADP
cana-6094	79	20	betti	betti	ADJ
cana-6094	79	21	numbers	number	NOUN
cana-6094	79	22	,	,	PUNCT
cana-6094	79	23	persistence	persistence	NOUN
cana-6094	79	24	diagrams	diagram	NOUN
cana-6094	79	25	,	,	PUNCT
cana-6094	79	26	and	and	CCONJ
cana-6094	79	27	homology	homology	NOUN
cana-6094	79	28	groups	group	NOUN
cana-6094	79	29	.	.	PUNCT
cana-6094	80	1	by	by	ADP
cana-6094	80	2	message	message	NOUN
cana-6094	80	3	passing	pass	VERB
cana-6094	80	4	along	along	ADP
cana-6094	80	5	face	face	NOUN
cana-6094	80	6	relations	relation	NOUN
cana-6094	80	7	,	,	PUNCT
cana-6094	80	8	gnns	gnns	PROPN
cana-6094	80	9	can	can	AUX
cana-6094	80	10	learn	learn	VERB
cana-6094	80	11	to	to	PART
cana-6094	80	12	approximate	approximate	VERB
cana-6094	80	13	topological	topological	ADJ
cana-6094	80	14	properties	property	NOUN
cana-6094	80	15	that	that	PRON
cana-6094	80	16	are	be	AUX
cana-6094	80	17	traditionally	traditionally	ADV
cana-6094	80	18	computed	compute	VERB
cana-6094	80	19	through	through	ADP
cana-6094	80	20	expensive	expensive	ADJ
cana-6094	80	21	algebraic	algebraic	ADJ
cana-6094	80	22	operations	operation	NOUN
cana-6094	80	23	8	8	NUM
cana-6094	80	24	.	.	PUNCT
cana-6094	81	1	this	this	DET
cana-6094	81	2	approach	approach	NOUN
cana-6094	81	3	has	have	VERB
cana-6094	81	4	applications	application	NOUN
cana-6094	81	5	in	in	ADP
cana-6094	81	6	computational	computational	ADJ
cana-6094	81	7	topology	topology	NOUN
cana-6094	81	8	and	and	CCONJ
cana-6094	81	9	shape	shape	VERB
cana-6094	81	10	analysis	analysis	NOUN
cana-6094	81	11	.	.	PUNCT
cana-6094	82	1	in	in	ADP
cana-6094	82	2	geometric	geometric	ADJ
cana-6094	82	3	deep	deep	ADJ
cana-6094	82	4	learning	learning	NOUN
cana-6094	82	5	,	,	PUNCT
cana-6094	82	6	equivariant	equivariant	ADJ
cana-6094	82	7	gnns	gnns	PROPN
cana-6094	82	8	preserve	preserve	VERB
cana-6094	82	9	geometric	geometric	ADJ
cana-6094	82	10	symmetries	symmetry	NOUN
cana-6094	82	11	while	while	SCONJ
cana-6094	82	12	analyzing	analyze	VERB
cana-6094	82	13	manifold	manifold	ADJ
cana-6094	82	14	structures	structure	NOUN
cana-6094	82	15	.	.	PUNCT
cana-6094	83	1	these	these	DET
cana-6094	83	2	networks	network	NOUN
cana-6094	83	3	have	have	AUX
cana-6094	83	4	been	be	AUX
cana-6094	83	5	applied	apply	VERB
cana-6094	83	6	to	to	ADP
cana-6094	83	7	problems	problem	NOUN
cana-6094	83	8	in	in	ADP
cana-6094	83	9	differential	differential	ADJ
cana-6094	83	10	geometry	geometry	NOUN
cana-6094	83	11	,	,	PUNCT
cana-6094	83	12	such	such	ADJ
cana-6094	83	13	as	as	ADP
cana-6094	83	14	estimating	estimate	VERB
cana-6094	83	15	curvature	curvature	NOUN
cana-6094	83	16	,	,	PUNCT
cana-6094	83	17	detecting	detect	VERB
cana-6094	83	18	geodesics	geodesic	NOUN
cana-6094	83	19	,	,	PUNCT
cana-6094	83	20	and	and	CCONJ
cana-6094	83	21	identifying	identify	VERB
cana-6094	83	22	isometries	isometry	NOUN
cana-6094	83	23	between	between	ADP
cana-6094	83	24	riemannian	riemannian	ADJ
cana-6094	83	25	manifolds	manifold	NOUN
cana-6094	83	26	8	8	NUM
cana-6094	83	27	.	.	PUNCT
cana-6094	83	28	by	by	ADP
cana-6094	83	29	preserving	preserve	VERB
cana-6094	83	30	geometric	geometric	ADJ
cana-6094	83	31	invariants	invariant	NOUN
cana-6094	83	32	,	,	PUNCT
cana-6094	83	33	these	these	DET
cana-6094	83	34	networks	network	NOUN
cana-6094	83	35	provide	provide	VERB
cana-6094	83	36	mathematically	mathematically	ADV
cana-6094	83	37	consistent	consistent	ADJ
cana-6094	83	38	analysis	analysis	NOUN
cana-6094	83	39	of	of	ADP
cana-6094	83	40	geometric	geometric	ADJ
cana-6094	83	41	structures	structure	NOUN
cana-6094	83	42	.	.	PUNCT
cana-6094	84	1	point	point	NOUN
cana-6094	84	2	cloud	cloud	NOUN
cana-6094	84	3	processing	processing	NOUN
cana-6094	84	4	represents	represent	VERB
cana-6094	84	5	another	another	DET
cana-6094	84	6	application	application	NOUN
cana-6094	84	7	where	where	SCONJ
cana-6094	84	8	gnns	gnns	ADJ
cana-6094	84	9	analyze	analyze	VERB
cana-6094	84	10	discrete	discrete	ADJ
cana-6094	84	11	samples	sample	NOUN
cana-6094	84	12	from	from	ADP
cana-6094	84	13	continuous	continuous	ADJ
cana-6094	84	14	geometric	geometric	ADJ
cana-6094	84	15	spaces	space	NOUN
cana-6094	84	16	.	.	PUNCT
cana-6094	85	1	by	by	ADP
cana-6094	85	2	constructing	construct	VERB
cana-6094	85	3	graphs	graph	NOUN
cana-6094	85	4	from	from	ADP
cana-6094	85	5	point	point	NOUN
cana-6094	85	6	neighborhoods	neighborhood	NOUN
cana-6094	85	7	,	,	PUNCT
cana-6094	85	8	gnns	gnns	PROPN
cana-6094	85	9	can	can	AUX
cana-6094	85	10	learn	learn	VERB
cana-6094	85	11	geometric	geometric	ADJ
cana-6094	85	12	properties	property	NOUN
cana-6094	85	13	and	and	CCONJ
cana-6094	85	14	classify	classify	VERB
cana-6094	85	15	shapes	shape	NOUN
cana-6094	85	16	based	base	VERB
cana-6094	85	17	on	on	ADP
cana-6094	85	18	their	their	PRON
cana-6094	85	19	structural	structural	ADJ
cana-6094	85	20	characteristics	characteristic	NOUN
cana-6094	85	21	9	9	NUM
cana-6094	85	22	.	.	PUNCT
cana-6094	86	1	this	this	DET
cana-6094	86	2	approach	approach	NOUN
cana-6094	86	3	bridges	bridge	NOUN
cana-6094	86	4	discrete	discrete	ADJ
cana-6094	86	5	graph	graph	NOUN
cana-6094	86	6	analysis	analysis	NOUN
cana-6094	86	7	with	with	ADP
cana-6094	86	8	continuous	continuous	ADJ
cana-6094	86	9	geometry	geometry	NOUN
cana-6094	86	10	.	.	PUNCT
cana-6094	87	1	4.3	4.3	NUM
cana-6094	87	2	combinatorial	combinatorial	ADJ
cana-6094	87	3	and	and	CCONJ
cana-6094	87	4	discrete	discrete	ADJ
cana-6094	87	5	structures	structure	NOUN
cana-6094	87	6	combinatorial	combinatorial	ADJ
cana-6094	87	7	structures	structure	NOUN
cana-6094	87	8	such	such	ADJ
cana-6094	87	9	as	as	ADP
cana-6094	87	10	graphs	graph	NOUN
cana-6094	87	11	themselves	themselves	PRON
cana-6094	87	12	,	,	PUNCT
cana-6094	87	13	matroids	matroid	NOUN
cana-6094	87	14	,	,	PUNCT
cana-6094	87	15	designs	design	NOUN
cana-6094	87	16	,	,	PUNCT
cana-6094	87	17	and	and	CCONJ
cana-6094	87	18	polytopes	polytopes	PROPN
cana-6094	87	19	naturally	naturally	ADV
cana-6094	87	20	admit	admit	VERB
cana-6094	87	21	graph	graph	NOUN
cana-6094	87	22	representations	representation	NOUN
cana-6094	87	23	that	that	SCONJ
cana-6094	87	24	gnns	gnns	NOUN
cana-6094	87	25	can	can	AUX
cana-6094	87	26	effectively	effectively	ADV
cana-6094	87	27	analyze	analyze	VERB
cana-6094	87	28	.	.	PUNCT
cana-6094	88	1	applications	application	NOUN
cana-6094	88	2	in	in	ADP
cana-6094	88	3	this	this	DET
cana-6094	88	4	domain	domain	NOUN
cana-6094	88	5	include	include	VERB
cana-6094	88	6	graph	graph	NOUN
cana-6094	88	7	property	property	NOUN
cana-6094	88	8	prediction	prediction	NOUN
cana-6094	88	9	,	,	PUNCT
cana-6094	88	10	combinatorial	combinatorial	ADJ
cana-6094	88	11	optimization	optimization	NOUN
cana-6094	88	12	,	,	PUNCT
cana-6094	88	13	and	and	CCONJ
cana-6094	88	14	structure	structure	NOUN
cana-6094	88	15	generation	generation	NOUN
cana-6094	88	16	19	19	NUM
cana-6094	88	17	.	.	PUNCT
cana-6094	89	1	for	for	ADP
cana-6094	89	2	graph	graph	NOUN
cana-6094	89	3	theory	theory	NOUN
cana-6094	89	4	,	,	PUNCT
cana-6094	89	5	gnns	gnns	PROPN
cana-6094	89	6	can	can	AUX
cana-6094	89	7	predict	predict	VERB
cana-6094	89	8	properties	property	NOUN
cana-6094	89	9	of	of	ADP
cana-6094	89	10	graphs	graph	NOUN
cana-6094	89	11	such	such	ADJ
cana-6094	89	12	as	as	ADP
cana-6094	89	13	connectivity	connectivity	NOUN
cana-6094	89	14	,	,	PUNCT
cana-6094	89	15	planarity	planarity	NOUN
cana-6094	89	16	,	,	PUNCT
cana-6094	89	17	colorability	colorability	NOUN
cana-6094	89	18	,	,	PUNCT
cana-6094	89	19	and	and	CCONJ
cana-6094	89	20	expansion	expansion	NOUN
cana-6094	89	21	properties	property	NOUN
cana-6094	89	22	.	.	PUNCT
cana-6094	90	1	by	by	ADP
cana-6094	90	2	learning	learn	VERB
cana-6094	90	3	from	from	ADP
cana-6094	90	4	structural	structural	ADJ
cana-6094	90	5	patterns	pattern	NOUN
cana-6094	90	6	,	,	PUNCT
cana-6094	90	7	gnns	gnn	NOUN
cana-6094	90	8	often	often	ADV
cana-6094	90	9	outperform	outperform	VERB
cana-6094	90	10	traditional	traditional	ADJ
cana-6094	90	11	heuristic	heuristic	ADJ
cana-6094	90	12	methods	method	NOUN
cana-6094	90	13	for	for	ADP
cana-6094	90	14	graph	graph	NOUN
cana-6094	90	15	property	property	NOUN
cana-6094	90	16	prediction	prediction	NOUN
cana-6094	90	17	,	,	PUNCT
cana-6094	90	18	particularly	particularly	ADV
cana-6094	90	19	on	on	ADP
cana-6094	90	20	large	large	ADJ
cana-6094	90	21	graphs	graph	NOUN
cana-6094	90	22	where	where	SCONJ
cana-6094	90	23	exhaustive	exhaustive	ADJ
cana-6094	90	24	analysis	analysis	NOUN
cana-6094	90	25	is	be	AUX
cana-6094	90	26	infeasible	infeasible	ADJ
cana-6094	90	27	19	19	NUM
cana-6094	90	28	.	.	PUNCT
cana-6094	91	1	in	in	ADP
cana-6094	91	2	combinatorial	combinatorial	ADJ
cana-6094	91	3	optimization	optimization	NOUN
cana-6094	91	4	,	,	PUNCT
cana-6094	91	5	gnns	gnn	NOUN
cana-6094	91	6	have	have	AUX
cana-6094	91	7	been	be	AUX
cana-6094	91	8	applied	apply	VERB
cana-6094	91	9	to	to	ADP
cana-6094	91	10	problems	problem	NOUN
cana-6094	91	11	such	such	ADJ
cana-6094	91	12	as	as	ADP
cana-6094	91	13	maximum	maximum	ADJ
cana-6094	91	14	cut	cut	NOUN
cana-6094	91	15	,	,	PUNCT
cana-6094	91	16	traveling	travel	VERB
cana-6094	91	17	salesman	salesman	NOUN
cana-6094	91	18	,	,	PUNCT
cana-6094	91	19	and	and	CCONJ
cana-6094	91	20	vertex	vertex	NOUN
cana-6094	91	21	cover	cover	NOUN
cana-6094	91	22	by	by	ADP
cana-6094	91	23	learning	learn	VERB
cana-6094	91	24	to	to	PART
cana-6094	91	25	approximate	approximate	VERB
cana-6094	91	26	solutions	solution	NOUN
cana-6094	91	27	based	base	VERB
cana-6094	91	28	on	on	ADP
cana-6094	91	29	graph	graph	NOUN
cana-6094	91	30	structure	structure	NOUN
cana-6094	91	31	.	.	PUNCT
cana-6094	92	1	these	these	DET
cana-6094	92	2	approaches	approach	NOUN
cana-6094	92	3	combine	combine	VERB
cana-6094	92	4	neural	neural	ADJ
cana-6094	92	5	network	network	NOUN
cana-6094	92	6	approximations	approximation	NOUN
cana-6094	92	7	with	with	ADP
cana-6094	92	8	traditional	traditional	ADJ
cana-6094	92	9	algorithms	algorithm	NOUN
cana-6094	92	10	to	to	PART
cana-6094	92	11	create	create	VERB
cana-6094	92	12	hybrid	hybrid	ADJ
cana-6094	92	13	methods	method	NOUN
cana-6094	92	14	that	that	PRON
cana-6094	92	15	leverage	leverage	VERB
cana-6094	92	16	both	both	PRON
cana-6094	92	17	learned	learn	VERB
cana-6094	92	18	patterns	pattern	NOUN
cana-6094	92	19	and	and	CCONJ
cana-6094	92	20	exact	exact	ADJ
cana-6094	92	21	computation	computation	NOUN
cana-6094	92	22	9	9	NUM
cana-6094	92	23	.	.	PUNCT
cana-6094	92	24	gnns	gnns	PROPN
cana-6094	92	25	have	have	AUX
cana-6094	92	26	also	also	ADV
cana-6094	92	27	been	be	AUX
cana-6094	92	28	used	use	VERB
cana-6094	92	29	for	for	ADP
cana-6094	92	30	combinatorial	combinatorial	ADJ
cana-6094	92	31	structure	structure	NOUN
cana-6094	92	32	generation	generation	NOUN
cana-6094	92	33	,	,	PUNCT
cana-6094	92	34	creating	create	VERB
cana-6094	92	35	new	new	ADJ
cana-6094	92	36	mathematical	mathematical	ADJ
cana-6094	92	37	objects	object	NOUN
cana-6094	92	38	with	with	ADP
cana-6094	92	39	desired	desire	VERB
cana-6094	92	40	properties	property	NOUN
cana-6094	92	41	.	.	PUNCT
cana-6094	93	1	for	for	ADP
cana-6094	93	2	example	example	NOUN
cana-6094	93	3	,	,	PUNCT
cana-6094	93	4	gnns	gnns	PROPN
cana-6094	93	5	can	can	AUX
cana-6094	93	6	generate	generate	VERB
cana-6094	93	7	graphs	graph	NOUN
cana-6094	93	8	with	with	ADP
cana-6094	93	9	specific	specific	ADJ
cana-6094	93	10	spectral	spectral	ADJ
cana-6094	93	11	properties	property	NOUN
cana-6094	93	12	,	,	PUNCT
cana-6094	93	13	designs	design	NOUN
cana-6094	93	14	with	with	ADP
cana-6094	93	15	particular	particular	ADJ
cana-6094	93	16	combinatorial	combinatorial	ADJ
cana-6094	93	17	parameters	parameter	NOUN
cana-6094	93	18	,	,	PUNCT
cana-6094	93	19	or	or	CCONJ
cana-6094	93	20	polytopes	polytope	NOUN
cana-6094	93	21	with	with	ADP
cana-6094	93	22	face	face	NOUN
cana-6094	93	23	numbers	number	NOUN
cana-6094	93	24	satisfying	satisfy	VERB
cana-6094	93	25	certain	certain	ADJ
cana-6094	93	26	inequalities	inequality	NOUN
cana-6094	93	27	9	9	NUM
cana-6094	93	28	.	.	PUNCT
cana-6094	94	1	this	this	DET
cana-6094	94	2	generative	generative	ADJ
cana-6094	94	3	capability	capability	NOUN
cana-6094	94	4	opens	open	VERB
cana-6094	94	5	new	new	ADJ
cana-6094	94	6	possibilities	possibility	NOUN
cana-6094	94	7	for	for	ADP
cana-6094	94	8	mathematical	mathematical	ADJ
cana-6094	94	9	discovery	discovery	NOUN
cana-6094	94	10	through	through	ADP
cana-6094	94	11	pattern	pattern	NOUN
cana-6094	94	12	learning	learning	NOUN
cana-6094	94	13	.	.	PUNCT
cana-6094	95	1	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	X
cana-6094	95	2	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	PROPN
cana-6094	95	3	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	PROPN
cana-6094	95	4	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	PRON
cana-6094	96	1	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	96	2	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	96	3	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	96	4	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	96	5	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	VERB
cana-6094	96	6	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	96	7	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	96	8	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	VERB
cana-6094	96	9	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	VERB
cana-6094	96	10	communications	communication	NOUN
cana-6094	96	11	on	on	ADP
cana-6094	96	12	applied	apply	VERB
cana-6094	96	13	nonlinear	nonlinear	ADJ
cana-6094	96	14	analysis	analysis	NOUN
cana-6094	96	15	issn	issn	NOUN
cana-6094	96	16	:	:	PUNCT
cana-6094	96	17	1074	1074	NUM
cana-6094	96	18	-	-	PUNCT
cana-6094	96	19	133x	133x	NUM
cana-6094	96	20	vol	vol	NOUN
cana-6094	96	21	29	29	NUM
cana-6094	96	22	no	no	NOUN
cana-6094	96	23	.	.	NOUN
cana-6094	96	24	4	4	NUM
cana-6094	96	25	(	(	PUNCT
cana-6094	96	26	2022	2022	NUM
cana-6094	96	27	)	)	PUNCT
cana-6094	96	28	33	33	NUM
cana-6094	96	29	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	96	30	4.4	4.4	NUM
cana-6094	96	31	mathematical	mathematical	ADJ
cana-6094	96	32	physics	physics	NOUN
cana-6094	96	33	and	and	CCONJ
cana-6094	96	34	complex	complex	ADJ
cana-6094	96	35	systems	system	NOUN
cana-6094	96	36	mathematical	mathematical	ADJ
cana-6094	96	37	physics	physics	NOUN
cana-6094	96	38	presents	present	VERB
cana-6094	96	39	rich	rich	ADJ
cana-6094	96	40	opportunities	opportunity	NOUN
cana-6094	96	41	for	for	ADP
cana-6094	96	42	gnn	gnn	PROPN
cana-6094	96	43	applications	application	NOUN
cana-6094	96	44	through	through	ADP
cana-6094	96	45	the	the	DET
cana-6094	96	46	analysis	analysis	NOUN
cana-6094	96	47	of	of	ADP
cana-6094	96	48	physical	physical	ADJ
cana-6094	96	49	systems	system	NOUN
cana-6094	96	50	represented	represent	VERB
cana-6094	96	51	as	as	ADP
cana-6094	96	52	graphs	graph	NOUN
cana-6094	96	53	.	.	PUNCT
cana-6094	97	1	from	from	ADP
cana-6094	97	2	molecular	molecular	ADJ
cana-6094	97	3	structures	structure	NOUN
cana-6094	97	4	to	to	ADP
cana-6094	97	5	spacetime	spacetime	NOUN
cana-6094	97	6	discretizations	discretization	NOUN
cana-6094	97	7	,	,	PUNCT
cana-6094	97	8	gnns	gnn	NOUN
cana-6094	97	9	provide	provide	VERB
cana-6094	97	10	powerful	powerful	ADJ
cana-6094	97	11	tools	tool	NOUN
cana-6094	97	12	for	for	ADP
cana-6094	97	13	learning	learn	VERB
cana-6094	97	14	physical	physical	ADJ
cana-6094	97	15	laws	law	NOUN
cana-6094	97	16	and	and	CCONJ
cana-6094	97	17	properties	property	NOUN
cana-6094	97	18	from	from	ADP
cana-6094	97	19	structural	structural	ADJ
cana-6094	97	20	data	datum	NOUN
cana-6094	97	21	89	89	NUM
cana-6094	97	22	.	.	PUNCT
cana-6094	98	1	in	in	ADP
cana-6094	98	2	quantum	quantum	ADJ
cana-6094	98	3	chemistry	chemistry	NOUN
cana-6094	98	4	,	,	PUNCT
cana-6094	98	5	gnns	gnn	NOUN
cana-6094	98	6	analyze	analyze	VERB
cana-6094	98	7	molecular	molecular	ADJ
cana-6094	98	8	graphs	graph	NOUN
cana-6094	98	9	to	to	PART
cana-6094	98	10	predict	predict	VERB
cana-6094	98	11	quantum	quantum	ADJ
cana-6094	98	12	mechanical	mechanical	ADJ
cana-6094	98	13	properties	property	NOUN
cana-6094	98	14	such	such	ADJ
cana-6094	98	15	as	as	ADP
cana-6094	98	16	energy	energy	NOUN
cana-6094	98	17	levels	level	NOUN
cana-6094	98	18	,	,	PUNCT
cana-6094	98	19	wave	wave	NOUN
cana-6094	98	20	functions	function	NOUN
cana-6094	98	21	,	,	PUNCT
cana-6094	98	22	and	and	CCONJ
cana-6094	98	23	reaction	reaction	NOUN
cana-6094	98	24	rates	rate	NOUN
cana-6094	98	25	.	.	PUNCT
cana-6094	99	1	equivariant	equivariant	PROPN
cana-6094	99	2	gnns	gnns	PROPN
cana-6094	99	3	particularly	particularly	ADV
cana-6094	99	4	excel	excel	VERB
cana-6094	99	5	in	in	ADP
cana-6094	99	6	this	this	DET
cana-6094	99	7	domain	domain	NOUN
cana-6094	99	8	by	by	ADP
cana-6094	99	9	preserving	preserve	VERB
cana-6094	99	10	physical	physical	ADJ
cana-6094	99	11	symmetries	symmetry	NOUN
cana-6094	99	12	such	such	ADJ
cana-6094	99	13	as	as	ADP
cana-6094	99	14	rotational	rotational	ADJ
cana-6094	99	15	invariance	invariance	NOUN
cana-6094	99	16	of	of	ADP
cana-6094	99	17	molecular	molecular	ADJ
cana-6094	99	18	properties	property	NOUN
cana-6094	99	19	8	8	NUM
cana-6094	99	20	.	.	PUNCT
cana-6094	100	1	these	these	DET
cana-6094	100	2	applications	application	NOUN
cana-6094	100	3	demonstrate	demonstrate	VERB
cana-6094	100	4	how	how	SCONJ
cana-6094	100	5	gnns	gnns	NOUN
cana-6094	100	6	can	can	AUX
cana-6094	100	7	learn	learn	VERB
cana-6094	100	8	complex	complex	ADJ
cana-6094	100	9	physical	physical	ADJ
cana-6094	100	10	relationships	relationship	NOUN
cana-6094	100	11	from	from	ADP
cana-6094	100	12	structural	structural	ADJ
cana-6094	100	13	data	datum	NOUN
cana-6094	100	14	.	.	PUNCT
cana-6094	101	1	for	for	ADP
cana-6094	101	2	general	general	ADJ
cana-6094	101	3	relativity	relativity	NOUN
cana-6094	101	4	and	and	CCONJ
cana-6094	101	5	cosmology	cosmology	NOUN
cana-6094	101	6	,	,	PUNCT
cana-6094	101	7	gnns	gnn	NOUN
cana-6094	101	8	have	have	AUX
cana-6094	101	9	been	be	AUX
cana-6094	101	10	applied	apply	VERB
cana-6094	101	11	to	to	ADP
cana-6094	101	12	discretized	discretize	VERB
cana-6094	101	13	representations	representation	NOUN
cana-6094	101	14	of	of	ADP
cana-6094	101	15	spacetime	spacetime	NOUN
cana-6094	101	16	to	to	PART
cana-6094	101	17	learn	learn	VERB
cana-6094	101	18	gravitational	gravitational	ADJ
cana-6094	101	19	properties	property	NOUN
cana-6094	101	20	,	,	PUNCT
cana-6094	101	21	detect	detect	NOUN
cana-6094	101	22	singularities	singularity	NOUN
cana-6094	101	23	,	,	PUNCT
cana-6094	101	24	and	and	CCONJ
cana-6094	101	25	identify	identify	VERB
cana-6094	101	26	cosmological	cosmological	ADJ
cana-6094	101	27	patterns	pattern	NOUN
cana-6094	101	28	.	.	PUNCT
cana-6094	102	1	by	by	ADP
cana-6094	102	2	representing	represent	VERB
cana-6094	102	3	spacetime	spacetime	NOUN
cana-6094	102	4	as	as	ADP
cana-6094	102	5	a	a	DET
cana-6094	102	6	graph	graph	NOUN
cana-6094	102	7	of	of	ADP
cana-6094	102	8	causal	causal	ADJ
cana-6094	102	9	relationships	relationship	NOUN
cana-6094	102	10	or	or	CCONJ
cana-6094	102	11	proximity	proximity	NOUN
cana-6094	102	12	connections	connection	NOUN
cana-6094	102	13	,	,	PUNCT
cana-6094	102	14	gnns	gnns	PROPN
cana-6094	102	15	can	can	AUX
cana-6094	102	16	learn	learn	VERB
cana-6094	102	17	to	to	PART
cana-6094	102	18	approximate	approximate	VERB
cana-6094	102	19	solutions	solution	NOUN
cana-6094	102	20	to	to	ADP
cana-6094	102	21	einstein	einstein	PROPN
cana-6094	102	22	's	's	PART
cana-6094	102	23	field	field	NOUN
cana-6094	102	24	equations	equation	NOUN
cana-6094	102	25	or	or	CCONJ
cana-6094	102	26	identify	identify	VERB
cana-6094	102	27	features	feature	NOUN
cana-6094	102	28	in	in	ADP
cana-6094	102	29	cosmological	cosmological	ADJ
cana-6094	102	30	datasets	dataset	NOUN
cana-6094	102	31	9	9	NUM
cana-6094	102	32	.	.	PUNCT
cana-6094	102	33	complex	complex	ADJ
cana-6094	102	34	systems	system	NOUN
cana-6094	102	35	across	across	ADP
cana-6094	102	36	scientific	scientific	ADJ
cana-6094	102	37	domains	domain	NOUN
cana-6094	102	38	often	often	ADV
cana-6094	102	39	admit	admit	VERB
cana-6094	102	40	graph	graph	NOUN
cana-6094	102	41	representations	representation	NOUN
cana-6094	102	42	that	that	SCONJ
cana-6094	102	43	gnns	gnns	NOUN
cana-6094	102	44	can	can	AUX
cana-6094	102	45	effectively	effectively	ADV
cana-6094	102	46	analyze	analyze	VERB
cana-6094	102	47	.	.	PUNCT
cana-6094	103	1	from	from	ADP
cana-6094	103	2	neural	neural	ADJ
cana-6094	103	3	connectivity	connectivity	NOUN
cana-6094	103	4	graphs	graph	NOUN
cana-6094	103	5	in	in	ADP
cana-6094	103	6	neuroscience	neuroscience	NOUN
cana-6094	103	7	to	to	ADP
cana-6094	103	8	energy	energy	NOUN
cana-6094	103	9	networks	network	NOUN
cana-6094	103	10	in	in	ADP
cana-6094	103	11	physics	physics	NOUN
cana-6094	103	12	,	,	PUNCT
cana-6094	103	13	gnns	gnns	PROPN
cana-6094	103	14	provide	provide	VERB
cana-6094	103	15	a	a	DET
cana-6094	103	16	unified	unified	ADJ
cana-6094	103	17	framework	framework	NOUN
cana-6094	103	18	for	for	ADP
cana-6094	103	19	extracting	extract	VERB
cana-6094	103	20	insights	insight	NOUN
cana-6094	103	21	from	from	ADP
cana-6094	103	22	diverse	diverse	ADJ
cana-6094	103	23	mathematical	mathematical	ADJ
cana-6094	103	24	structures	structure	NOUN
cana-6094	103	25	representing	represent	VERB
cana-6094	103	26	complex	complex	ADJ
cana-6094	103	27	systems	system	NOUN
cana-6094	103	28	19	19	NUM
cana-6094	103	29	.	.	PUNCT
cana-6094	104	1	mathematical	mathematical	ADJ
cana-6094	104	2	domain	domain	NOUN
cana-6094	104	3	key	key	NOUN
cana-6094	104	4	applications	application	NOUN
cana-6094	104	5	gnn	gnn	NOUN
cana-6094	104	6	strengths	strength	NOUN
cana-6094	104	7	notable	notable	ADJ
cana-6094	104	8	achievements	achievement	NOUN
cana-6094	104	9	algebraic	algebraic	ADJ
cana-6094	104	10	structures	structure	NOUN
cana-6094	104	11	group	group	NOUN
cana-6094	104	12	property	property	NOUN
cana-6094	104	13	prediction	prediction	NOUN
cana-6094	104	14	,	,	PUNCT
cana-6094	104	15	theorem	theorem	VERB
cana-6094	104	16	proving	prove	VERB
cana-6094	104	17	pattern	pattern	NOUN
cana-6094	104	18	recognition	recognition	NOUN
cana-6094	104	19	in	in	ADP
cana-6094	104	20	cayley	cayley	ADJ
cana-6094	104	21	graphs	graph	NOUN
cana-6094	104	22	predicting	predict	VERB
cana-6094	104	23	group	group	NOUN
cana-6094	104	24	properties	property	NOUN
cana-6094	104	25	from	from	ADP
cana-6094	104	26	generator	generator	NOUN
cana-6094	104	27	relations	relation	NOUN
cana-6094	104	28	geometry	geometry	NOUN
cana-6094	104	29	/	/	SYM
cana-6094	104	30	topology	topology	NOUN
cana-6094	104	31	shape	shape	NOUN
cana-6094	104	32	analysis	analysis	NOUN
cana-6094	104	33	,	,	PUNCT
cana-6094	104	34	topological	topological	ADJ
cana-6094	104	35	invariants	invariant	NOUN
cana-6094	104	36	processing	process	VERB
cana-6094	104	37	discretized	discretize	VERB
cana-6094	104	38	manifolds	manifold	NOUN
cana-6094	104	39	,	,	PUNCT
cana-6094	104	40	simplicial	simplicial	ADJ
cana-6094	104	41	complexes	complex	NOUN
cana-6094	104	42	approximating	approximate	VERB
cana-6094	104	43	betti	betti	NOUN
cana-6094	104	44	numbers	number	NOUN
cana-6094	104	45	,	,	PUNCT
cana-6094	104	46	curvature	curvature	NOUN
cana-6094	104	47	estimation	estimation	NOUN
cana-6094	104	48	combinatorics	combinatoric	NOUN
cana-6094	104	49	graph	graph	NOUN
cana-6094	104	50	property	property	NOUN
cana-6094	104	51	prediction	prediction	NOUN
cana-6094	104	52	,	,	PUNCT
cana-6094	104	53	optimization	optimization	NOUN
cana-6094	104	54	structure	structure	NOUN
cana-6094	104	55	learning	learn	VERB
cana-6094	104	56	on	on	ADP
cana-6094	104	57	combinatorial	combinatorial	ADJ
cana-6094	104	58	graphs	graph	NOUN
cana-6094	104	59	solving	solve	VERB
cana-6094	104	60	combinatorial	combinatorial	ADJ
cana-6094	104	61	problems	problem	NOUN
cana-6094	104	62	,	,	PUNCT
cana-6094	104	63	generating	generate	VERB
cana-6094	104	64	designs	design	NOUN
cana-6094	104	65	mathematical	mathematical	ADJ
cana-6094	104	66	physics	physics	NOUN
cana-6094	104	67	molecular	molecular	ADJ
cana-6094	104	68	analysis	analysis	NOUN
cana-6094	104	69	,	,	PUNCT
cana-6094	104	70	spacetime	spacetime	NOUN
cana-6094	104	71	modeling	modeling	NOUN
cana-6094	104	72	symmetry	symmetry	NOUN
cana-6094	104	73	preservation	preservation	NOUN
cana-6094	104	74	,	,	PUNCT
cana-6094	104	75	physical	physical	ADJ
cana-6094	104	76	invariance	invariance	NOUN
cana-6094	104	77	predicting	predict	VERB
cana-6094	104	78	quantum	quantum	ADJ
cana-6094	104	79	properties	property	NOUN
cana-6094	104	80	,	,	PUNCT
cana-6094	104	81	gravitational	gravitational	ADJ
cana-6094	104	82	analysis	analysis	NOUN
cana-6094	104	83	number	number	NOUN
cana-6094	104	84	theory	theory	NOUN
cana-6094	104	85	prime	prime	ADJ
cana-6094	104	86	distribution	distribution	NOUN
cana-6094	104	87	,	,	PUNCT
cana-6094	104	88	integer	integer	NOUN
cana-6094	104	89	sequences	sequence	NOUN
cana-6094	104	90	pattern	pattern	NOUN
cana-6094	104	91	detection	detection	NOUN
cana-6094	104	92	in	in	ADP
cana-6094	104	93	numerical	numerical	ADJ
cana-6094	104	94	graphs	graph	NOUN
cana-6094	104	95	identifying	identify	VERB
cana-6094	104	96	number	number	NOUN
cana-6094	104	97	patterns	pattern	NOUN
cana-6094	104	98	,	,	PUNCT
cana-6094	104	99	conjecture	conjecture	NOUN
cana-6094	104	100	generation	generation	NOUN
cana-6094	104	101	table	table	NOUN
cana-6094	104	102	3	3	NUM
cana-6094	104	103	:	:	PUNCT
cana-6094	104	104	gnn	gnn	NOUN
cana-6094	104	105	applications	application	NOUN
cana-6094	104	106	across	across	ADP
cana-6094	104	107	mathematical	mathematical	ADJ
cana-6094	104	108	domains	domain	NOUN
cana-6094	104	109	5.challenges	5.challenges	NUM
cana-6094	104	110	and	and	CCONJ
cana-6094	104	111	limitations	limitation	NOUN
cana-6094	104	112	5.1	5.1	NUM
cana-6094	104	113	theoretical	theoretical	ADJ
cana-6094	104	114	guarantees	guarantee	NOUN
cana-6094	104	115	and	and	CCONJ
cana-6094	104	116	mathematical	mathematical	ADJ
cana-6094	104	117	rigor	rigor	NOUN
cana-6094	104	118	despite	despite	SCONJ
cana-6094	104	119	their	their	PRON
cana-6094	104	120	empirical	empirical	ADJ
cana-6094	104	121	successes	success	NOUN
cana-6094	104	122	,	,	PUNCT
cana-6094	104	123	gnns	gnns	ADJ
cana-6094	104	124	face	face	VERB
cana-6094	104	125	significant	significant	ADJ
cana-6094	104	126	challenges	challenge	NOUN
cana-6094	104	127	in	in	ADP
cana-6094	104	128	providing	provide	VERB
cana-6094	104	129	theoretical	theoretical	ADJ
cana-6094	104	130	guarantees	guarantee	NOUN
cana-6094	104	131	that	that	PRON
cana-6094	104	132	meet	meet	VERB
cana-6094	104	133	the	the	DET
cana-6094	104	134	standards	standard	NOUN
cana-6094	104	135	of	of	ADP
cana-6094	104	136	mathematical	mathematical	ADJ
cana-6094	104	137	rigor	rigor	NOUN
cana-6094	104	138	.	.	PUNCT
cana-6094	105	1	unlike	unlike	ADP
cana-6094	105	2	traditional	traditional	ADJ
cana-6094	105	3	mathematical	mathematical	ADJ
cana-6094	105	4	methods	method	NOUN
cana-6094	105	5	that	that	PRON
cana-6094	105	6	offer	offer	VERB
cana-6094	105	7	formal	formal	ADJ
cana-6094	105	8	proofs	proof	NOUN
cana-6094	105	9	and	and	CCONJ
cana-6094	105	10	error	error	NOUN
cana-6094	105	11	bounds	bound	NOUN
cana-6094	105	12	,	,	PUNCT
cana-6094	105	13	gnns	gnn	NOUN
cana-6094	105	14	typically	typically	ADV
cana-6094	105	15	provide	provide	VERB
cana-6094	105	16	statistical	statistical	ADJ
cana-6094	105	17	approximations	approximation	NOUN
cana-6094	105	18	without	without	ADP
cana-6094	105	19	certainty	certainty	NOUN
cana-6094	105	20	guarantees	guarantee	NOUN
cana-6094	105	21	10	10	NUM
cana-6094	105	22	.	.	PUNCT
cana-6094	106	1	this	this	DET
cana-6094	106	2	limitation	limitation	NOUN
cana-6094	106	3	is	be	AUX
cana-6094	106	4	particularly	particularly	ADV
cana-6094	106	5	problematic	problematic	ADJ
cana-6094	106	6	in	in	ADP
cana-6094	106	7	mathematical	mathematical	ADJ
cana-6094	106	8	contexts	contexts	NOUN
cana-6094	106	9	where	where	SCONJ
cana-6094	106	10	exact	exact	ADJ
cana-6094	106	11	results	result	NOUN
cana-6094	106	12	are	be	AUX
cana-6094	106	13	required	require	VERB
cana-6094	106	14	.	.	PUNCT
cana-6094	107	1	the	the	DET
cana-6094	107	2	generalization	generalization	NOUN
cana-6094	107	3	properties	property	NOUN
cana-6094	107	4	of	of	ADP
cana-6094	107	5	gnns	gnns	NOUN
cana-6094	107	6	on	on	ADP
cana-6094	107	7	mathematical	mathematical	ADJ
cana-6094	107	8	structures	structure	NOUN
cana-6094	107	9	are	be	AUX
cana-6094	107	10	not	not	PART
cana-6094	107	11	well	well	ADV
cana-6094	107	12	understood	understand	VERB
cana-6094	107	13	theoretically	theoretically	ADV
cana-6094	107	14	.	.	PUNCT
cana-6094	108	1	while	while	SCONJ
cana-6094	108	2	these	these	DET
cana-6094	108	3	models	model	NOUN
cana-6094	108	4	often	often	ADV
cana-6094	108	5	perform	perform	VERB
cana-6094	108	6	well	well	ADV
cana-6094	108	7	on	on	ADP
cana-6094	108	8	training	training	NOUN
cana-6094	108	9	distributions	distribution	NOUN
cana-6094	108	10	,	,	PUNCT
cana-6094	108	11	they	they	PRON
cana-6094	108	12	may	may	AUX
cana-6094	108	13	fail	fail	VERB
cana-6094	108	14	on	on	ADP
cana-6094	108	15	out	out	ADV
cana-6094	108	16	-	-	PUNCT
cana-6094	108	17	of	of	ADP
cana-6094	108	18	-	-	PUNCT
cana-6094	108	19	distribution	distribution	NOUN
cana-6094	108	20	mathematical	mathematical	ADJ
cana-6094	108	21	objects	object	NOUN
cana-6094	108	22	or	or	CCONJ
cana-6094	108	23	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NUM
cana-6094	108	24	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	108	25	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	108	26	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	109	1	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	109	2	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	109	3	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	109	4	communications	communication	NOUN
cana-6094	109	5	on	on	ADP
cana-6094	109	6	applied	apply	VERB
cana-6094	109	7	nonlinear	nonlinear	ADJ
cana-6094	109	8	analysis	analysis	NOUN
cana-6094	109	9	issn	issn	NOUN
cana-6094	109	10	:	:	PUNCT
cana-6094	109	11	1074	1074	NUM
cana-6094	109	12	-	-	PUNCT
cana-6094	109	13	133x	133x	NUM
cana-6094	109	14	vol	vol	NOUN
cana-6094	109	15	29	29	NUM
cana-6094	109	16	no	no	NOUN
cana-6094	109	17	.	.	NOUN
cana-6094	109	18	4	4	NUM
cana-6094	109	19	(	(	PUNCT
cana-6094	109	20	2022	2022	NUM
cana-6094	109	21	)	)	PUNCT
cana-6094	109	22	34	34	NUM
cana-6094	109	23	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	109	24	edge	edge	NOUN
cana-6094	109	25	cases	case	NOUN
cana-6094	109	26	that	that	PRON
cana-6094	109	27	deviate	deviate	VERB
cana-6094	109	28	from	from	ADP
cana-6094	109	29	training	training	NOUN
cana-6094	109	30	patterns	pattern	NOUN
cana-6094	109	31	10	10	NUM
cana-6094	109	32	.	.	PUNCT
cana-6094	110	1	this	this	DET
cana-6094	110	2	limitation	limitation	NOUN
cana-6094	110	3	necessitates	necessitate	VERB
cana-6094	110	4	careful	careful	ADJ
cana-6094	110	5	validation	validation	NOUN
cana-6094	110	6	through	through	ADP
cana-6094	110	7	traditional	traditional	ADJ
cana-6094	110	8	mathematical	mathematical	ADJ
cana-6094	110	9	methods	method	NOUN
cana-6094	110	10	when	when	SCONJ
cana-6094	110	11	applying	apply	VERB
cana-6094	110	12	gnns	gnn	NOUN
cana-6094	110	13	to	to	ADP
cana-6094	110	14	mathematical	mathematical	ADJ
cana-6094	110	15	discovery	discovery	NOUN
cana-6094	110	16	.	.	PUNCT
cana-6094	111	1	expressivity	expressivity	NOUN
cana-6094	111	2	limits	limit	NOUN
cana-6094	111	3	represent	represent	VERB
cana-6094	111	4	another	another	DET
cana-6094	111	5	theoretical	theoretical	ADJ
cana-6094	111	6	challenge	challenge	NOUN
cana-6094	111	7	.	.	PUNCT
cana-6094	112	1	while	while	SCONJ
cana-6094	112	2	gnns	gnn	NOUN
cana-6094	112	3	can	can	AUX
cana-6094	112	4	capture	capture	VERB
cana-6094	112	5	many	many	ADJ
cana-6094	112	6	graph	graph	NOUN
cana-6094	112	7	patterns	pattern	NOUN
cana-6094	112	8	,	,	PUNCT
cana-6094	112	9	they	they	PRON
cana-6094	112	10	have	have	AUX
cana-6094	112	11	known	know	VERB
cana-6094	112	12	limitations	limitation	NOUN
cana-6094	112	13	in	in	ADP
cana-6094	112	14	distinguishing	distinguish	VERB
cana-6094	112	15	certain	certain	ADJ
cana-6094	112	16	graph	graph	NOUN
cana-6094	112	17	isomorphism	isomorphism	NOUN
cana-6094	112	18	classes	class	NOUN
cana-6094	112	19	or	or	CCONJ
cana-6094	112	20	capturing	capture	VERB
cana-6094	112	21	higher	high	ADJ
cana-6094	112	22	-	-	PUNCT
cana-6094	112	23	order	order	NOUN
cana-6094	112	24	logical	logical	ADJ
cana-6094	112	25	properties	property	NOUN
cana-6094	112	26	1	1	NUM
cana-6094	112	27	.	.	PUNCT
cana-6094	113	1	these	these	DET
cana-6094	113	2	expressivity	expressivity	NOUN
cana-6094	113	3	constraints	constraint	NOUN
cana-6094	113	4	may	may	AUX
cana-6094	113	5	limit	limit	VERB
cana-6094	113	6	their	their	PRON
cana-6094	113	7	ability	ability	NOUN
cana-6094	113	8	to	to	PART
cana-6094	113	9	capture	capture	VERB
cana-6094	113	10	sophisticated	sophisticated	ADJ
cana-6094	113	11	mathematical	mathematical	ADJ
cana-6094	113	12	properties	property	NOUN
cana-6094	113	13	that	that	PRON
cana-6094	113	14	require	require	VERB
cana-6094	113	15	reasoning	reasoning	NOUN
cana-6094	113	16	beyond	beyond	ADP
cana-6094	113	17	local	local	ADJ
cana-6094	113	18	graph	graph	NOUN
cana-6094	113	19	patterns	pattern	NOUN
cana-6094	113	20	.	.	PUNCT
cana-6094	114	1	5.2	5.2	NUM
cana-6094	114	2	data	datum	NOUN
cana-6094	114	3	scarcity	scarcity	NOUN
cana-6094	114	4	and	and	CCONJ
cana-6094	114	5	representation	representation	NOUN
cana-6094	114	6	issues	issue	NOUN
cana-6094	114	7	data	datum	NOUN
cana-6094	114	8	scarcity	scarcity	NOUN
cana-6094	114	9	presents	present	VERB
cana-6094	114	10	a	a	DET
cana-6094	114	11	significant	significant	ADJ
cana-6094	114	12	challenge	challenge	NOUN
cana-6094	114	13	for	for	ADP
cana-6094	114	14	applying	apply	VERB
cana-6094	114	15	gnns	gnn	NOUN
cana-6094	114	16	to	to	ADP
cana-6094	114	17	mathematical	mathematical	ADJ
cana-6094	114	18	structures	structure	NOUN
cana-6094	114	19	,	,	PUNCT
cana-6094	114	20	particularly	particularly	ADV
cana-6094	114	21	for	for	ADP
cana-6094	114	22	specialized	specialized	ADJ
cana-6094	114	23	or	or	CCONJ
cana-6094	114	24	novel	novel	ADJ
cana-6094	114	25	mathematical	mathematical	ADJ
cana-6094	114	26	objects	object	NOUN
cana-6094	114	27	where	where	SCONJ
cana-6094	114	28	few	few	ADJ
cana-6094	114	29	examples	example	NOUN
cana-6094	114	30	exist	exist	VERB
cana-6094	114	31	for	for	ADP
cana-6094	114	32	training	training	NOUN
cana-6094	114	33	.	.	PUNCT
cana-6094	115	1	unlike	unlike	ADP
cana-6094	115	2	domains	domain	NOUN
cana-6094	115	3	like	like	ADP
cana-6094	115	4	social	social	ADJ
cana-6094	115	5	networks	network	NOUN
cana-6094	115	6	or	or	CCONJ
cana-6094	115	7	molecular	molecular	ADJ
cana-6094	115	8	biology	biology	NOUN
cana-6094	115	9	where	where	SCONJ
cana-6094	115	10	large	large	ADJ
cana-6094	115	11	datasets	dataset	NOUN
cana-6094	115	12	are	be	AUX
cana-6094	115	13	available	available	ADJ
cana-6094	115	14	,	,	PUNCT
cana-6094	115	15	many	many	ADJ
cana-6094	115	16	mathematical	mathematical	ADJ
cana-6094	115	17	areas	area	NOUN
cana-6094	115	18	lack	lack	VERB
cana-6094	115	19	sufficient	sufficient	ADJ
cana-6094	115	20	data	datum	NOUN
cana-6094	115	21	for	for	ADP
cana-6094	115	22	effective	effective	ADJ
cana-6094	115	23	machine	machine	NOUN
cana-6094	115	24	learning	learn	VERB
cana-6094	115	25	10	10	NUM
cana-6094	115	26	.	.	PUNCT
cana-6094	116	1	the	the	DET
cana-6094	116	2	representation	representation	NOUN
cana-6094	116	3	problem	problem	NOUN
cana-6094	116	4	—	—	PUNCT
cana-6094	116	5	how	how	SCONJ
cana-6094	116	6	to	to	PART
cana-6094	116	7	best	well	ADV
cana-6094	116	8	represent	represent	VERB
cana-6094	116	9	mathematical	mathematical	ADJ
cana-6094	116	10	structures	structure	NOUN
cana-6094	116	11	as	as	ADP
cana-6094	116	12	graphs	graph	NOUN
cana-6094	116	13	—	—	PUNCT
cana-6094	116	14	introduces	introduce	VERB
cana-6094	116	15	another	another	DET
cana-6094	116	16	layer	layer	NOUN
cana-6094	116	17	of	of	ADP
cana-6094	116	18	complexity	complexity	NOUN
cana-6094	116	19	.	.	PUNCT
cana-6094	117	1	different	different	ADJ
cana-6094	117	2	graph	graph	NOUN
cana-6094	117	3	representations	representation	NOUN
cana-6094	117	4	of	of	ADP
cana-6094	117	5	the	the	DET
cana-6094	117	6	same	same	ADJ
cana-6094	117	7	mathematical	mathematical	ADJ
cana-6094	117	8	object	object	NOUN
cana-6094	117	9	may	may	AUX
cana-6094	117	10	emphasize	emphasize	VERB
cana-6094	117	11	different	different	ADJ
cana-6094	117	12	properties	property	NOUN
cana-6094	117	13	,	,	PUNCT
cana-6094	117	14	leading	lead	VERB
cana-6094	117	15	to	to	ADP
cana-6094	117	16	varying	vary	VERB
cana-6094	117	17	gnn	gnn	PROPN
cana-6094	117	18	performance	performance	NOUN
cana-6094	117	19	7	7	NUM
cana-6094	117	20	.	.	PUNCT
cana-6094	117	21	choosing	choose	VERB
cana-6094	117	22	appropriate	appropriate	ADJ
cana-6094	117	23	representations	representation	NOUN
cana-6094	117	24	requires	require	VERB
cana-6094	117	25	mathematical	mathematical	ADJ
cana-6094	117	26	insight	insight	NOUN
cana-6094	117	27	that	that	PRON
cana-6094	117	28	may	may	AUX
cana-6094	117	29	defeat	defeat	VERB
cana-6094	117	30	the	the	DET
cana-6094	117	31	purpose	purpose	NOUN
cana-6094	117	32	of	of	ADP
cana-6094	117	33	automated	automate	VERB
cana-6094	117	34	analysis	analysis	NOUN
cana-6094	117	35	.	.	PUNCT
cana-6094	118	1	invariance	invariance	NOUN
cana-6094	118	2	and	and	CCONJ
cana-6094	118	3	equivariance	equivariance	NOUN
cana-6094	118	4	requirements	requirement	NOUN
cana-6094	118	5	vary	vary	VERB
cana-6094	118	6	across	across	ADP
cana-6094	118	7	mathematical	mathematical	ADJ
cana-6094	118	8	contexts	contexts	NOUN
cana-6094	118	9	,	,	PUNCT
cana-6094	118	10	and	and	CCONJ
cana-6094	118	11	designing	designing	NOUN
cana-6094	118	12	gnns	gnn	NOUN
cana-6094	118	13	that	that	PRON
cana-6094	118	14	preserve	preserve	VERB
cana-6094	118	15	all	all	DET
cana-6094	118	16	relevant	relevant	ADJ
cana-6094	118	17	symmetries	symmetry	NOUN
cana-6094	118	18	can	can	AUX
cana-6094	118	19	be	be	AUX
cana-6094	118	20	challenging	challenge	VERB
cana-6094	118	21	.	.	PUNCT
cana-6094	119	1	while	while	SCONJ
cana-6094	119	2	specialized	specialized	ADJ
cana-6094	119	3	architectures	architecture	NOUN
cana-6094	119	4	like	like	ADP
cana-6094	119	5	equivariant	equivariant	PROPN
cana-6094	119	6	gnns	gnns	PROPN
cana-6094	119	7	address	address	NOUN
cana-6094	119	8	some	some	DET
cana-6094	119	9	symmetry	symmetry	NOUN
cana-6094	119	10	requirements	requirement	NOUN
cana-6094	119	11	,	,	PUNCT
cana-6094	119	12	no	no	DET
cana-6094	119	13	universal	universal	ADJ
cana-6094	119	14	framework	framework	NOUN
cana-6094	119	15	exists	exist	VERB
cana-6094	119	16	for	for	ADP
cana-6094	119	17	preserving	preserve	VERB
cana-6094	119	18	all	all	DET
cana-6094	119	19	mathematical	mathematical	ADJ
cana-6094	119	20	invariants	invariant	NOUN
cana-6094	119	21	that	that	PRON
cana-6094	119	22	might	might	AUX
cana-6094	119	23	be	be	AUX
cana-6094	119	24	relevant	relevant	ADJ
cana-6094	119	25	across	across	ADP
cana-6094	119	26	different	different	ADJ
cana-6094	119	27	mathematical	mathematical	ADJ
cana-6094	119	28	domains	domain	NOUN
cana-6094	119	29	8	8	NUM
cana-6094	119	30	.	.	X
cana-6094	119	31	5.3	5.3	NUM
cana-6094	119	32	computational	computational	ADJ
cana-6094	119	33	complexity	complexity	NOUN
cana-6094	119	34	and	and	CCONJ
cana-6094	119	35	scalability	scalability	NOUN
cana-6094	119	36	computational	computational	ADJ
cana-6094	119	37	complexity	complexity	NOUN
cana-6094	119	38	limits	limit	VERB
cana-6094	119	39	the	the	DET
cana-6094	119	40	application	application	NOUN
cana-6094	119	41	of	of	ADP
cana-6094	119	42	gnns	gnns	NOUN
cana-6094	119	43	to	to	ADP
cana-6094	119	44	large	large	ADJ
cana-6094	119	45	mathematical	mathematical	ADJ
cana-6094	119	46	structures	structure	NOUN
cana-6094	119	47	.	.	PUNCT
cana-6094	120	1	while	while	SCONJ
cana-6094	120	2	gnns	gnn	NOUN
cana-6094	120	3	are	be	AUX
cana-6094	120	4	generally	generally	ADV
cana-6094	120	5	efficient	efficient	ADJ
cana-6094	120	6	for	for	ADP
cana-6094	120	7	graphs	graph	NOUN
cana-6094	120	8	of	of	ADP
cana-6094	120	9	moderate	moderate	ADJ
cana-6094	120	10	size	size	NOUN
cana-6094	120	11	,	,	PUNCT
cana-6094	120	12	extremely	extremely	ADV
cana-6094	120	13	large	large	ADJ
cana-6094	120	14	mathematical	mathematical	ADJ
cana-6094	120	15	objects	object	NOUN
cana-6094	120	16	(	(	PUNCT
cana-6094	120	17	e.g.	e.g.	ADV
cana-6094	120	18	,	,	PUNCT
cana-6094	120	19	massive	massive	ADJ
cana-6094	120	20	groups	group	NOUN
cana-6094	120	21	,	,	PUNCT
cana-6094	120	22	highdimensional	highdimensional	NOUN
cana-6094	120	23	manifolds	manifold	NOUN
cana-6094	120	24	,	,	PUNCT
cana-6094	120	25	or	or	CCONJ
cana-6094	120	26	complex	complex	ADJ
cana-6094	120	27	combinatorial	combinatorial	ADJ
cana-6094	120	28	structures	structure	NOUN
cana-6094	120	29	)	)	PUNCT
cana-6094	120	30	may	may	AUX
cana-6094	120	31	present	present	VERB
cana-6094	120	32	computational	computational	ADJ
cana-6094	120	33	challenges	challenge	NOUN
cana-6094	120	34	14	14	NUM
cana-6094	120	35	.	.	PUNCT
cana-6094	121	1	the	the	DET
cana-6094	121	2	message	message	NOUN
cana-6094	121	3	passing	pass	VERB
cana-6094	121	4	mechanism	mechanism	NOUN
cana-6094	121	5	inherently	inherently	ADV
cana-6094	121	6	limits	limit	VERB
cana-6094	121	7	the	the	DET
cana-6094	121	8	receptive	receptive	ADJ
cana-6094	121	9	field	field	NOUN
cana-6094	121	10	of	of	ADP
cana-6094	121	11	each	each	DET
cana-6094	121	12	node	node	NOUN
cana-6094	121	13	to	to	ADP
cana-6094	121	14	its	its	PRON
cana-6094	121	15	neighborhood	neighborhood	NOUN
cana-6094	121	16	,	,	PUNCT
cana-6094	121	17	requiring	require	VERB
cana-6094	121	18	multiple	multiple	ADJ
cana-6094	121	19	layers	layer	NOUN
cana-6094	121	20	to	to	PART
cana-6094	121	21	capture	capture	VERB
cana-6094	121	22	global	global	ADJ
cana-6094	121	23	properties	property	NOUN
cana-6094	121	24	.	.	PUNCT
cana-6094	122	1	for	for	ADP
cana-6094	122	2	structures	structure	NOUN
cana-6094	122	3	with	with	ADP
cana-6094	122	4	large	large	ADJ
cana-6094	122	5	diameters	diameter	NOUN
cana-6094	122	6	or	or	CCONJ
cana-6094	122	7	long	long	ADJ
cana-6094	122	8	-	-	PUNCT
cana-6094	122	9	range	range	NOUN
cana-6094	122	10	dependencies	dependency	NOUN
cana-6094	122	11	,	,	PUNCT
cana-6094	122	12	this	this	PRON
cana-6094	122	13	may	may	AUX
cana-6094	122	14	necessitate	necessitate	VERB
cana-6094	122	15	impractically	impractically	ADV
cana-6094	122	16	deep	deep	ADJ
cana-6094	122	17	networks	network	NOUN
cana-6094	122	18	or	or	CCONJ
cana-6094	122	19	specialized	specialized	ADJ
cana-6094	122	20	architectures	architecture	NOUN
cana-6094	122	21	7	7	NUM
cana-6094	122	22	.	.	PUNCT
cana-6094	123	1	this	this	DET
cana-6094	123	2	limitation	limitation	NOUN
cana-6094	123	3	is	be	AUX
cana-6094	123	4	particularly	particularly	ADV
cana-6094	123	5	relevant	relevant	ADJ
cana-6094	123	6	for	for	ADP
cana-6094	123	7	mathematical	mathematical	ADJ
cana-6094	123	8	structures	structure	NOUN
cana-6094	123	9	where	where	SCONJ
cana-6094	123	10	global	global	ADJ
cana-6094	123	11	properties	property	NOUN
cana-6094	123	12	emerge	emerge	VERB
cana-6094	123	13	from	from	ADP
cana-6094	123	14	intricate	intricate	ADJ
cana-6094	123	15	long	long	ADJ
cana-6094	123	16	-	-	PUNCT
cana-6094	123	17	range	range	NOUN
cana-6094	123	18	relationships	relationship	NOUN
cana-6094	123	19	.	.	PUNCT
cana-6094	124	1	memory	memory	NOUN
cana-6094	124	2	requirements	requirement	NOUN
cana-6094	124	3	for	for	ADP
cana-6094	124	4	storing	store	VERB
cana-6094	124	5	and	and	CCONJ
cana-6094	124	6	processing	process	VERB
cana-6094	124	7	large	large	ADJ
cana-6094	124	8	mathematical	mathematical	ADJ
cana-6094	124	9	graphs	graph	NOUN
cana-6094	124	10	can	can	AUX
cana-6094	124	11	be	be	AUX
cana-6094	124	12	prohibitive	prohibitive	ADJ
cana-6094	124	13	,	,	PUNCT
cana-6094	124	14	especially	especially	ADV
cana-6094	124	15	when	when	SCONJ
cana-6094	124	16	node	node	ADJ
cana-6094	124	17	features	feature	NOUN
cana-6094	124	18	are	be	AUX
cana-6094	124	19	high	high	ADV
cana-6094	124	20	-	-	PUNCT
cana-6094	124	21	dimensional	dimensional	ADJ
cana-6094	124	22	or	or	CCONJ
cana-6094	124	23	when	when	SCONJ
cana-6094	124	24	the	the	DET
cana-6094	124	25	graphs	graph	NOUN
cana-6094	124	26	are	be	AUX
cana-6094	124	27	densely	densely	ADV
cana-6094	124	28	connected	connect	VERB
cana-6094	124	29	.	.	PUNCT
cana-6094	125	1	these	these	DET
cana-6094	125	2	constraints	constraint	NOUN
cana-6094	125	3	may	may	AUX
cana-6094	125	4	limit	limit	VERB
cana-6094	125	5	the	the	DET
cana-6094	125	6	size	size	NOUN
cana-6094	125	7	and	and	CCONJ
cana-6094	125	8	complexity	complexity	NOUN
cana-6094	125	9	of	of	ADP
cana-6094	125	10	mathematical	mathematical	ADJ
cana-6094	125	11	structures	structure	NOUN
cana-6094	125	12	that	that	PRON
cana-6094	125	13	can	can	AUX
cana-6094	125	14	be	be	AUX
cana-6094	125	15	practically	practically	ADV
cana-6094	125	16	analyzed	analyze	VERB
cana-6094	125	17	with	with	ADP
cana-6094	125	18	gnns	gnns	ADJ
cana-6094	125	19	4	4	NUM
cana-6094	125	20	.	.	NOUN
cana-6094	125	21	5.4	5.4	NUM
cana-6094	125	22	interpretability	interpretability	NOUN
cana-6094	125	23	and	and	CCONJ
cana-6094	125	24	mathematical	mathematical	ADJ
cana-6094	125	25	insight	insight	NOUN
cana-6094	125	26	interpretability	interpretability	NOUN
cana-6094	125	27	challenges	challenge	NOUN
cana-6094	125	28	represent	represent	VERB
cana-6094	125	29	a	a	DET
cana-6094	125	30	significant	significant	ADJ
cana-6094	125	31	limitation	limitation	NOUN
cana-6094	125	32	for	for	ADP
cana-6094	125	33	mathematical	mathematical	ADJ
cana-6094	125	34	applications	application	NOUN
cana-6094	125	35	where	where	SCONJ
cana-6094	125	36	understanding	understand	VERB
cana-6094	125	37	why	why	SCONJ
cana-6094	125	38	a	a	DET
cana-6094	125	39	model	model	NOUN
cana-6094	125	40	makes	make	VERB
cana-6094	125	41	certain	certain	ADJ
cana-6094	125	42	predictions	prediction	NOUN
cana-6094	125	43	is	be	AUX
cana-6094	125	44	as	as	ADV
cana-6094	125	45	important	important	ADJ
cana-6094	125	46	as	as	ADP
cana-6094	125	47	the	the	DET
cana-6094	125	48	predictions	prediction	NOUN
cana-6094	125	49	themselves	themselves	PRON
cana-6094	125	50	.	.	PUNCT
cana-6094	126	1	while	while	SCONJ
cana-6094	126	2	gnns	gnn	NOUN
cana-6094	126	3	can	can	AUX
cana-6094	126	4	identify	identify	VERB
cana-6094	126	5	patterns	pattern	NOUN
cana-6094	126	6	and	and	CCONJ
cana-6094	126	7	make	make	VERB
cana-6094	126	8	accurate	accurate	ADJ
cana-6094	126	9	predictions	prediction	NOUN
cana-6094	126	10	,	,	PUNCT
cana-6094	126	11	they	they	PRON
cana-6094	126	12	often	often	ADV
cana-6094	126	13	function	function	VERB
cana-6094	126	14	as	as	ADP
cana-6094	126	15	black	black	ADJ
cana-6094	126	16	boxes	box	NOUN
cana-6094	126	17	that	that	PRON
cana-6094	126	18	provide	provide	VERB
cana-6094	126	19	limited	limited	ADJ
cana-6094	126	20	insight	insight	NOUN
cana-6094	126	21	into	into	ADP
cana-6094	126	22	the	the	DET
cana-6094	126	23	underlying	underlying	ADJ
cana-6094	126	24	mathematical	mathematical	ADJ
cana-6094	126	25	reasons	reason	NOUN
cana-6094	126	26	for	for	ADP
cana-6094	126	27	their	their	PRON
cana-6094	126	28	predictions	prediction	NOUN
cana-6094	126	29	310	310	NUM
cana-6094	126	30	.	.	PUNCT
cana-6094	127	1	the	the	DET
cana-6094	127	2	correlation	correlation	NOUN
cana-6094	127	3	vs.	vs.	ADP
cana-6094	127	4	causation	causation	NOUN
cana-6094	127	5	dilemma	dilemma	NOUN
cana-6094	127	6	is	be	AUX
cana-6094	127	7	particularly	particularly	ADV
cana-6094	127	8	acute	acute	ADJ
cana-6094	127	9	in	in	ADP
cana-6094	127	10	mathematical	mathematical	ADJ
cana-6094	127	11	contexts	contexts	NOUN
cana-6094	127	12	where	where	SCONJ
cana-6094	127	13	causal	causal	ADJ
cana-6094	127	14	relationships	relationship	NOUN
cana-6094	127	15	are	be	AUX
cana-6094	127	16	fundamental	fundamental	ADJ
cana-6094	127	17	.	.	PUNCT
cana-6094	128	1	gnns	gnns	PROPN
cana-6094	128	2	may	may	AUX
cana-6094	128	3	identify	identify	VERB
cana-6094	128	4	spurious	spurious	ADJ
cana-6094	128	5	correlations	correlation	NOUN
cana-6094	128	6	that	that	PRON
cana-6094	128	7	hold	hold	VERB
cana-6094	128	8	for	for	ADP
cana-6094	128	9	training	training	NOUN
cana-6094	128	10	data	datum	NOUN
cana-6094	128	11	but	but	CCONJ
cana-6094	128	12	fail	fail	VERB
cana-6094	128	13	to	to	PART
cana-6094	128	14	capture	capture	VERB
cana-6094	128	15	genuine	genuine	ADJ
cana-6094	128	16	mathematical	mathematical	ADJ
cana-6094	128	17	causality	causality	NOUN
cana-6094	128	18	,	,	PUNCT
cana-6094	128	19	potentially	potentially	ADV
cana-6094	128	20	leading	lead	VERB
cana-6094	128	21	to	to	ADP
cana-6094	128	22	incorrect	incorrect	ADJ
cana-6094	128	23	conjectures	conjecture	NOUN
cana-6094	128	24	or	or	CCONJ
cana-6094	128	25	proofs	proof	NOUN
cana-6094	128	26	10	10	NUM
cana-6094	128	27	.	.	PUNCT
cana-6094	129	1	integration	integration	NOUN
cana-6094	129	2	with	with	ADP
cana-6094	129	3	formal	formal	ADJ
cana-6094	129	4	methods	method	NOUN
cana-6094	129	5	remains	remain	VERB
cana-6094	129	6	challenging	challenging	ADJ
cana-6094	129	7	,	,	PUNCT
cana-6094	129	8	as	as	SCONJ
cana-6094	129	9	gnn	gnn	PROPN
cana-6094	129	10	predictions	prediction	NOUN
cana-6094	129	11	typically	typically	ADV
cana-6094	129	12	require	require	VERB
cana-6094	129	13	validation	validation	NOUN
cana-6094	129	14	through	through	ADP
cana-6094	129	15	traditional	traditional	ADJ
cana-6094	129	16	mathematical	mathematical	ADJ
cana-6094	129	17	reasoning	reasoning	NOUN
cana-6094	129	18	.	.	PUNCT
cana-6094	130	1	developing	develop	VERB
cana-6094	130	2	frameworks	framework	NOUN
cana-6094	130	3	that	that	PRON
cana-6094	130	4	seamlessly	seamlessly	ADV
cana-6094	130	5	combine	combine	VERB
cana-6094	130	6	gnn	gnn	PROPN
cana-6094	130	7	pattern	pattern	NOUN
cana-6094	130	8	recognition	recognition	NOUN
cana-6094	130	9	with	with	ADP
cana-6094	130	10	formal	formal	ADJ
cana-6094	130	11	verification	verification	NOUN
cana-6094	130	12	represents	represent	VERB
cana-6094	130	13	an	an	DET
cana-6094	130	14	important	important	ADJ
cana-6094	130	15	ongoing	ongoing	ADJ
cana-6094	130	16	research	research	NOUN
cana-6094	130	17	direction	direction	NOUN
cana-6094	130	18	910	910	NUM
cana-6094	130	19	.	.	PUNCT
cana-6094	131	1	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	131	2	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	131	3	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	131	4	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	X
cana-6094	131	5	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	SYM
cana-6094	131	6	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	131	7	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	NOUN
cana-6094	131	8	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	VERB
cana-6094	131	9	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-applications	https://neptune.ai/blog/graph-neural-network-and-some-of-gnn-application	NOUN
cana-6094	131	10	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	https://www.sciencedirect.com/science/article/abs/pii/s0925231224015327	NOUN
cana-6094	131	11	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	131	12	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	NOUN
cana-6094	131	13	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	PROPN
cana-6094	131	14	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	131	15	communications	communication	NOUN
cana-6094	131	16	on	on	ADP
cana-6094	131	17	applied	apply	VERB
cana-6094	131	18	nonlinear	nonlinear	ADJ
cana-6094	131	19	analysis	analysis	NOUN
cana-6094	131	20	issn	issn	NOUN
cana-6094	131	21	:	:	PUNCT
cana-6094	131	22	1074	1074	NUM
cana-6094	131	23	-	-	PUNCT
cana-6094	131	24	133x	133x	NUM
cana-6094	131	25	vol	vol	NOUN
cana-6094	131	26	29	29	NUM
cana-6094	131	27	no	no	NOUN
cana-6094	131	28	.	.	NOUN
cana-6094	131	29	4	4	NUM
cana-6094	131	30	(	(	PUNCT
cana-6094	131	31	2022	2022	NUM
cana-6094	131	32	)	)	PUNCT
cana-6094	131	33	35	35	NUM
cana-6094	131	34	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	131	35	6.future	6.future	NUM
cana-6094	131	36	directions	direction	NOUN
cana-6094	131	37	6.1	6.1	NUM
cana-6094	131	38	integration	integration	NOUN
cana-6094	131	39	with	with	ADP
cana-6094	131	40	symbolic	symbolic	ADJ
cana-6094	131	41	ai	ai	NOUN
cana-6094	131	42	and	and	CCONJ
cana-6094	131	43	theorem	theorem	VERB
cana-6094	131	44	proving	prove	VERB
cana-6094	131	45	the	the	DET
cana-6094	131	46	integration	integration	NOUN
cana-6094	131	47	of	of	ADP
cana-6094	131	48	gnns	gnns	NOUN
cana-6094	131	49	with	with	ADP
cana-6094	131	50	symbolic	symbolic	ADJ
cana-6094	131	51	ai	ai	PROPN
cana-6094	131	52	and	and	CCONJ
cana-6094	131	53	automated	automate	VERB
cana-6094	131	54	theorem	theorem	ADJ
cana-6094	131	55	proving	proving	NOUN
cana-6094	131	56	represents	represent	VERB
cana-6094	131	57	a	a	DET
cana-6094	131	58	promising	promising	ADJ
cana-6094	131	59	direction	direction	NOUN
cana-6094	131	60	for	for	ADP
cana-6094	131	61	enhancing	enhance	VERB
cana-6094	131	62	mathematical	mathematical	ADJ
cana-6094	131	63	reasoning	reasoning	NOUN
cana-6094	131	64	.	.	PUNCT
cana-6094	132	1	by	by	ADP
cana-6094	132	2	combining	combine	VERB
cana-6094	132	3	gnns	gnn	NOUN
cana-6094	132	4	'	'	PART
cana-6094	132	5	pattern	pattern	NOUN
cana-6094	132	6	recognition	recognition	NOUN
cana-6094	132	7	capabilities	capability	NOUN
cana-6094	132	8	with	with	ADP
cana-6094	132	9	symbolic	symbolic	ADJ
cana-6094	132	10	systems	system	NOUN
cana-6094	132	11	'	'	PART
cana-6094	132	12	logical	logical	ADJ
cana-6094	132	13	rigor	rigor	NOUN
cana-6094	132	14	,	,	PUNCT
cana-6094	132	15	researchers	researcher	NOUN
cana-6094	132	16	can	can	AUX
cana-6094	132	17	develop	develop	VERB
cana-6094	132	18	hybrid	hybrid	ADJ
cana-6094	132	19	approaches	approach	NOUN
cana-6094	132	20	that	that	PRON
cana-6094	132	21	leverage	leverage	VERB
cana-6094	132	22	the	the	DET
cana-6094	132	23	strengths	strength	NOUN
cana-6094	132	24	of	of	ADP
cana-6094	132	25	both	both	PRON
cana-6094	132	26	paradigms	paradigm	NOUN
cana-6094	132	27	9	9	NUM
cana-6094	132	28	.	.	PUNCT
cana-6094	133	1	future	future	ADJ
cana-6094	133	2	work	work	NOUN
cana-6094	133	3	may	may	AUX
cana-6094	133	4	focus	focus	VERB
cana-6094	133	5	on	on	ADP
cana-6094	133	6	using	use	VERB
cana-6094	133	7	gnns	gnn	NOUN
cana-6094	133	8	to	to	PART
cana-6094	133	9	guide	guide	VERB
cana-6094	133	10	proof	proof	NOUN
cana-6094	133	11	search	search	NOUN
cana-6094	133	12	in	in	ADP
cana-6094	133	13	theorem	theorem	ADJ
cana-6094	133	14	provers	prover	NOUN
cana-6094	133	15	by	by	ADP
cana-6094	133	16	suggesting	suggest	VERB
cana-6094	133	17	promising	promising	ADJ
cana-6094	133	18	proof	proof	NOUN
cana-6094	133	19	steps	step	NOUN
cana-6094	133	20	based	base	VERB
cana-6094	133	21	on	on	ADP
cana-6094	133	22	structural	structural	ADJ
cana-6094	133	23	patterns	pattern	NOUN
cana-6094	133	24	.	.	PUNCT
cana-6094	134	1	gnns	gnns	PROPN
cana-6094	134	2	could	could	AUX
cana-6094	134	3	learn	learn	VERB
cana-6094	134	4	to	to	PART
cana-6094	134	5	predict	predict	VERB
cana-6094	134	6	which	which	DET
cana-6094	134	7	proof	proof	NOUN
cana-6094	134	8	tactics	tactic	NOUN
cana-6094	134	9	or	or	CCONJ
cana-6094	134	10	lemmas	lemma	NOUN
cana-6094	134	11	are	be	AUX
cana-6094	134	12	likely	likely	ADJ
cana-6094	134	13	to	to	PART
cana-6094	134	14	be	be	AUX
cana-6094	134	15	effective	effective	ADJ
cana-6094	134	16	for	for	ADP
cana-6094	134	17	proving	prove	VERB
cana-6094	134	18	specific	specific	ADJ
cana-6094	134	19	mathematical	mathematical	ADJ
cana-6094	134	20	statements	statement	NOUN
cana-6094	134	21	,	,	PUNCT
cana-6094	134	22	reducing	reduce	VERB
cana-6094	134	23	the	the	DET
cana-6094	134	24	search	search	NOUN
cana-6094	134	25	space	space	NOUN
cana-6094	134	26	for	for	ADP
cana-6094	134	27	automated	automate	VERB
cana-6094	134	28	provers	prover	NOUN
cana-6094	134	29	9	9	NUM
cana-6094	134	30	.	.	PUNCT
cana-6094	135	1	another	another	DET
cana-6094	135	2	direction	direction	NOUN
cana-6094	135	3	involves	involve	VERB
cana-6094	135	4	using	use	VERB
cana-6094	135	5	gnns	gnn	NOUN
cana-6094	135	6	for	for	ADP
cana-6094	135	7	conjecture	conjecture	NOUN
cana-6094	135	8	generation	generation	NOUN
cana-6094	135	9	by	by	ADP
cana-6094	135	10	identifying	identify	VERB
cana-6094	135	11	patterns	pattern	NOUN
cana-6094	135	12	in	in	ADP
cana-6094	135	13	mathematical	mathematical	ADJ
cana-6094	135	14	structures	structure	NOUN
cana-6094	135	15	that	that	PRON
cana-6094	135	16	may	may	AUX
cana-6094	135	17	indicate	indicate	VERB
cana-6094	135	18	previously	previously	ADV
cana-6094	135	19	unnoticed	unnoticed	ADJ
cana-6094	135	20	relationships	relationship	NOUN
cana-6094	135	21	or	or	CCONJ
cana-6094	135	22	properties	property	NOUN
cana-6094	135	23	.	.	PUNCT
cana-6094	136	1	these	these	DET
cana-6094	136	2	statistically	statistically	ADV
cana-6094	136	3	-	-	PUNCT
cana-6094	136	4	generated	generate	VERB
cana-6094	136	5	conjectures	conjecture	NOUN
cana-6094	136	6	could	could	AUX
cana-6094	136	7	then	then	ADV
cana-6094	136	8	be	be	AUX
cana-6094	136	9	verified	verify	VERB
cana-6094	136	10	formally	formally	ADV
cana-6094	136	11	,	,	PUNCT
cana-6094	136	12	potentially	potentially	ADV
cana-6094	136	13	accelerating	accelerate	VERB
cana-6094	136	14	mathematical	mathematical	ADJ
cana-6094	136	15	discovery	discovery	NOUN
cana-6094	136	16	9	9	NUM
cana-6094	136	17	.	.	X
cana-6094	137	1	6.2	6.2	NUM
cana-6094	137	2	advanced	advance	VERB
cana-6094	137	3	geometrical	geometrical	ADJ
cana-6094	137	4	and	and	CCONJ
cana-6094	137	5	topological	topological	ADJ
cana-6094	137	6	learning	learn	VERB
cana-6094	137	7	advanced	advance	VERB
cana-6094	137	8	geometrical	geometrical	ADJ
cana-6094	137	9	learning	learning	NOUN
cana-6094	137	10	techniques	technique	NOUN
cana-6094	137	11	will	will	AUX
cana-6094	137	12	likely	likely	ADV
cana-6094	137	13	focus	focus	VERB
cana-6094	137	14	on	on	ADP
cana-6094	137	15	developing	develop	VERB
cana-6094	137	16	more	more	ADV
cana-6094	137	17	sophisticated	sophisticated	ADJ
cana-6094	137	18	equivariant	equivariant	ADJ
cana-6094	137	19	and	and	CCONJ
cana-6094	137	20	invariant	invariant	ADJ
cana-6094	137	21	gnns	gnn	NOUN
cana-6094	137	22	that	that	PRON
cana-6094	137	23	preserve	preserve	VERB
cana-6094	137	24	broader	broad	ADJ
cana-6094	137	25	classes	class	NOUN
cana-6094	137	26	of	of	ADP
cana-6094	137	27	mathematical	mathematical	ADJ
cana-6094	137	28	symmetries	symmetry	NOUN
cana-6094	137	29	.	.	PUNCT
cana-6094	138	1	future	future	ADJ
cana-6094	138	2	architectures	architecture	NOUN
cana-6094	138	3	may	may	AUX
cana-6094	138	4	automatically	automatically	ADV
cana-6094	138	5	discover	discover	VERB
cana-6094	138	6	relevant	relevant	ADJ
cana-6094	138	7	symmetries	symmetry	NOUN
cana-6094	138	8	from	from	ADP
cana-6094	138	9	data	datum	NOUN
cana-6094	138	10	rather	rather	ADV
cana-6094	138	11	than	than	ADP
cana-6094	138	12	requiring	require	VERB
cana-6094	138	13	manual	manual	ADJ
cana-6094	138	14	specification	specification	NOUN
cana-6094	138	15	of	of	ADP
cana-6094	138	16	symmetry	symmetry	NOUN
cana-6094	138	17	groups	group	NOUN
cana-6094	138	18	8	8	NUM
cana-6094	138	19	.	.	PUNCT
cana-6094	139	1	for	for	ADP
cana-6094	139	2	topological	topological	ADJ
cana-6094	139	3	data	datum	NOUN
cana-6094	139	4	analysis	analysis	NOUN
cana-6094	139	5	,	,	PUNCT
cana-6094	139	6	future	future	ADJ
cana-6094	139	7	work	work	NOUN
cana-6094	139	8	may	may	AUX
cana-6094	139	9	integrate	integrate	VERB
cana-6094	139	10	gnns	gnns	NOUN
cana-6094	139	11	with	with	ADP
cana-6094	139	12	persistent	persistent	ADJ
cana-6094	139	13	homology	homology	NOUN
cana-6094	139	14	and	and	CCONJ
cana-6094	139	15	other	other	ADJ
cana-6094	139	16	topological	topological	ADJ
cana-6094	139	17	methods	method	NOUN
cana-6094	139	18	to	to	PART
cana-6094	139	19	create	create	VERB
cana-6094	139	20	unified	unified	ADJ
cana-6094	139	21	frameworks	framework	NOUN
cana-6094	139	22	for	for	ADP
cana-6094	139	23	learning	learn	VERB
cana-6094	139	24	from	from	ADP
cana-6094	139	25	both	both	CCONJ
cana-6094	139	26	discrete	discrete	ADJ
cana-6094	139	27	graph	graph	NOUN
cana-6094	139	28	structure	structure	NOUN
cana-6094	139	29	and	and	CCONJ
cana-6094	139	30	continuous	continuous	ADJ
cana-6094	139	31	topological	topological	ADJ
cana-6094	139	32	properties	property	NOUN
cana-6094	139	33	.	.	PUNCT
cana-6094	140	1	these	these	DET
cana-6094	140	2	integrations	integration	NOUN
cana-6094	140	3	could	could	AUX
cana-6094	140	4	enable	enable	VERB
cana-6094	140	5	more	more	ADJ
cana-6094	140	6	powerful	powerful	ADJ
cana-6094	140	7	analysis	analysis	NOUN
cana-6094	140	8	of	of	ADP
cana-6094	140	9	shapes	shape	NOUN
cana-6094	140	10	and	and	CCONJ
cana-6094	140	11	spaces	space	NOUN
cana-6094	140	12	8	8	NUM
cana-6094	140	13	.	.	PUNCT
cana-6094	141	1	multi	multi	ADJ
cana-6094	141	2	-	-	ADJ
cana-6094	141	3	scale	scale	ADJ
cana-6094	141	4	geometric	geometric	ADJ
cana-6094	141	5	analysis	analysis	NOUN
cana-6094	141	6	represents	represent	VERB
cana-6094	141	7	another	another	DET
cana-6094	141	8	promising	promising	ADJ
cana-6094	141	9	direction	direction	NOUN
cana-6094	141	10	,	,	PUNCT
cana-6094	141	11	where	where	SCONJ
cana-6094	141	12	gnns	gnns	NOUN
cana-6094	141	13	would	would	AUX
cana-6094	141	14	operate	operate	VERB
cana-6094	141	15	simultaneously	simultaneously	ADV
cana-6094	141	16	at	at	ADP
cana-6094	141	17	multiple	multiple	ADJ
cana-6094	141	18	scales	scale	NOUN
cana-6094	141	19	of	of	ADP
cana-6094	141	20	geometric	geometric	ADJ
cana-6094	141	21	organization	organization	NOUN
cana-6094	141	22	.	.	PUNCT
cana-6094	142	1	such	such	ADJ
cana-6094	142	2	approaches	approach	NOUN
cana-6094	142	3	could	could	AUX
cana-6094	142	4	capture	capture	VERB
cana-6094	142	5	both	both	PRON
cana-6094	142	6	local	local	ADJ
cana-6094	142	7	geometric	geometric	ADJ
cana-6094	142	8	details	detail	NOUN
cana-6094	142	9	and	and	CCONJ
cana-6094	142	10	global	global	ADJ
cana-6094	142	11	geometric	geometric	ADJ
cana-6094	142	12	properties	property	NOUN
cana-6094	142	13	,	,	PUNCT
cana-6094	142	14	mirroring	mirror	VERB
cana-6094	142	15	how	how	SCONJ
cana-6094	142	16	mathematicians	mathematician	NOUN
cana-6094	142	17	analyze	analyze	VERB
cana-6094	142	18	geometric	geometric	ADJ
cana-6094	142	19	structures	structure	NOUN
cana-6094	142	20	8	8	NUM
cana-6094	142	21	.	.	X
cana-6094	142	22	6.3	6.3	NUM
cana-6094	142	23	cross	cross	ADJ
cana-6094	142	24	-	-	ADJ
cana-6094	142	25	domain	domain	ADJ
cana-6094	142	26	mathematical	mathematical	ADJ
cana-6094	142	27	applications	application	NOUN
cana-6094	142	28	cross	cross	ADJ
cana-6094	142	29	-	-	ADJ
cana-6094	142	30	domain	domain	ADJ
cana-6094	142	31	applications	application	NOUN
cana-6094	142	32	will	will	AUX
cana-6094	142	33	likely	likely	ADV
cana-6094	142	34	expand	expand	VERB
cana-6094	142	35	as	as	SCONJ
cana-6094	142	36	gnn	gnn	PROPN
cana-6094	142	37	methodologies	methodology	NOUN
cana-6094	142	38	mature	mature	VERB
cana-6094	142	39	,	,	PUNCT
cana-6094	142	40	applying	apply	VERB
cana-6094	142	41	similar	similar	ADJ
cana-6094	142	42	techniques	technique	NOUN
cana-6094	142	43	across	across	ADP
cana-6094	142	44	diverse	diverse	ADJ
cana-6094	142	45	mathematical	mathematical	ADJ
cana-6094	142	46	domains	domain	NOUN
cana-6094	142	47	that	that	PRON
cana-6094	142	48	share	share	VERB
cana-6094	142	49	structural	structural	ADJ
cana-6094	142	50	similarities	similarity	NOUN
cana-6094	142	51	.	.	PUNCT
cana-6094	143	1	for	for	ADP
cana-6094	143	2	example	example	NOUN
cana-6094	143	3	,	,	PUNCT
cana-6094	143	4	approaches	approach	NOUN
cana-6094	143	5	developed	develop	VERB
cana-6094	143	6	for	for	ADP
cana-6094	143	7	group	group	NOUN
cana-6094	143	8	theory	theory	NOUN
cana-6094	143	9	might	might	AUX
cana-6094	143	10	be	be	AUX
cana-6094	143	11	adapted	adapt	VERB
cana-6094	143	12	to	to	AUX
cana-6094	143	13	ring	ring	NOUN
cana-6094	143	14	theory	theory	NOUN
cana-6094	143	15	or	or	CCONJ
cana-6094	143	16	topology	topology	NOUN
cana-6094	143	17	through	through	ADP
cana-6094	143	18	appropriate	appropriate	ADJ
cana-6094	143	19	graph	graph	NOUN
cana-6094	143	20	representations	representation	NOUN
cana-6094	143	21	7	7	NUM
cana-6094	143	22	.	.	NOUN
cana-6094	143	23	mathematical	mathematical	ADJ
cana-6094	143	24	knowledge	knowledge	NOUN
cana-6094	143	25	graphs	graph	NOUN
cana-6094	143	26	that	that	PRON
cana-6094	143	27	integrate	integrate	VERB
cana-6094	143	28	concepts	concept	NOUN
cana-6094	143	29	from	from	ADP
cana-6094	143	30	multiple	multiple	ADJ
cana-6094	143	31	mathematical	mathematical	ADJ
cana-6094	143	32	domains	domain	NOUN
cana-6094	143	33	could	could	AUX
cana-6094	143	34	provide	provide	VERB
cana-6094	143	35	rich	rich	ADJ
cana-6094	143	36	training	training	NOUN
cana-6094	143	37	data	datum	NOUN
cana-6094	143	38	for	for	ADP
cana-6094	143	39	gnns	gnn	NOUN
cana-6094	143	40	to	to	PART
cana-6094	143	41	learn	learn	VERB
cana-6094	143	42	cross	cross	ADJ
cana-6094	143	43	-	-	ADJ
cana-6094	143	44	domain	domain	ADJ
cana-6094	143	45	mathematical	mathematical	ADJ
cana-6094	143	46	patterns	pattern	NOUN
cana-6094	143	47	.	.	PUNCT
cana-6094	144	1	these	these	DET
cana-6094	144	2	knowledge	knowledge	NOUN
cana-6094	144	3	graphs	graph	NOUN
cana-6094	144	4	would	would	AUX
cana-6094	144	5	represent	represent	VERB
cana-6094	144	6	mathematical	mathematical	ADJ
cana-6094	144	7	concepts	concept	NOUN
cana-6094	144	8	as	as	ADP
cana-6094	144	9	nodes	node	NOUN
cana-6094	144	10	and	and	CCONJ
cana-6094	144	11	relationships	relationship	NOUN
cana-6094	144	12	(	(	PUNCT
cana-6094	144	13	e.g.	e.g.	ADV
cana-6094	144	14	,	,	PUNCT
cana-6094	144	15	implication	implication	NOUN
cana-6094	144	16	,	,	PUNCT
cana-6094	144	17	generalization	generalization	NOUN
cana-6094	144	18	,	,	PUNCT
cana-6094	144	19	analogy	analogy	NOUN
cana-6094	144	20	)	)	PUNCT
cana-6094	144	21	as	as	ADP
cana-6094	144	22	edges	edge	NOUN
cana-6094	144	23	,	,	PUNCT
cana-6094	144	24	enabling	enable	VERB
cana-6094	144	25	gnns	gnn	NOUN
cana-6094	144	26	to	to	PART
cana-6094	144	27	discover	discover	VERB
cana-6094	144	28	connections	connection	NOUN
cana-6094	144	29	across	across	ADP
cana-6094	144	30	different	different	ADJ
cana-6094	144	31	mathematical	mathematical	ADJ
cana-6094	144	32	areas	area	NOUN
cana-6094	144	33	9	9	NUM
cana-6094	144	34	.	.	PUNCT
cana-6094	145	1	domain	domain	NOUN
cana-6094	145	2	adaptation	adaptation	NOUN
cana-6094	145	3	techniques	technique	NOUN
cana-6094	145	4	specifically	specifically	ADV
cana-6094	145	5	designed	design	VERB
cana-6094	145	6	for	for	ADP
cana-6094	145	7	mathematical	mathematical	ADJ
cana-6094	145	8	contexts	contexts	NOUN
cana-6094	145	9	will	will	AUX
cana-6094	145	10	be	be	AUX
cana-6094	145	11	important	important	ADJ
cana-6094	145	12	for	for	ADP
cana-6094	145	13	transferring	transfer	VERB
cana-6094	145	14	insights	insight	NOUN
cana-6094	145	15	from	from	ADP
cana-6094	145	16	data	data	NOUN
cana-6094	145	17	-	-	PUNCT
cana-6094	145	18	rich	rich	ADJ
cana-6094	145	19	mathematical	mathematical	ADJ
cana-6094	145	20	domains	domain	NOUN
cana-6094	145	21	to	to	ADP
cana-6094	145	22	data	data	NOUN
cana-6094	145	23	-	-	PUNCT
cana-6094	145	24	scarce	scarce	ADJ
cana-6094	145	25	ones	one	NOUN
cana-6094	145	26	.	.	PUNCT
cana-6094	146	1	these	these	DET
cana-6094	146	2	techniques	technique	NOUN
cana-6094	146	3	would	would	AUX
cana-6094	146	4	allow	allow	VERB
cana-6094	146	5	models	model	NOUN
cana-6094	146	6	trained	train	VERB
cana-6094	146	7	on	on	ADP
cana-6094	146	8	well	well	ADV
cana-6094	146	9	-	-	PUNCT
cana-6094	146	10	studied	study	VERB
cana-6094	146	11	mathematical	mathematical	ADJ
cana-6094	146	12	structures	structure	NOUN
cana-6094	146	13	to	to	PART
cana-6094	146	14	inform	inform	VERB
cana-6094	146	15	analysis	analysis	NOUN
cana-6094	146	16	of	of	ADP
cana-6094	146	17	less	less	ADV
cana-6094	146	18	-	-	PUNCT
cana-6094	146	19	studied	study	VERB
cana-6094	146	20	areas	area	NOUN
cana-6094	146	21	10	10	NUM
cana-6094	146	22	.	.	PUNCT
cana-6094	146	23	6.4	6.4	NUM
cana-6094	146	24	theoretical	theoretical	ADJ
cana-6094	146	25	advances	advance	NOUN
cana-6094	146	26	and	and	CCONJ
cana-6094	146	27	mathematical	mathematical	ADJ
cana-6094	146	28	guarantees	guarantee	NOUN
cana-6094	146	29	theoretical	theoretical	ADJ
cana-6094	146	30	advances	advance	NOUN
cana-6094	146	31	in	in	ADP
cana-6094	146	32	understanding	understand	VERB
cana-6094	146	33	gnn	gnn	NOUN
cana-6094	146	34	expressivity	expressivity	NOUN
cana-6094	146	35	and	and	CCONJ
cana-6094	146	36	limitations	limitation	NOUN
cana-6094	146	37	will	will	AUX
cana-6094	146	38	be	be	AUX
cana-6094	146	39	crucial	crucial	ADJ
cana-6094	146	40	for	for	ADP
cana-6094	146	41	their	their	PRON
cana-6094	146	42	mathematical	mathematical	ADJ
cana-6094	146	43	applications	application	NOUN
cana-6094	146	44	.	.	PUNCT
cana-6094	147	1	future	future	ADJ
cana-6094	147	2	work	work	NOUN
cana-6094	147	3	may	may	AUX
cana-6094	147	4	develop	develop	VERB
cana-6094	147	5	gnn	gnn	PROPN
cana-6094	147	6	architectures	architecture	NOUN
cana-6094	147	7	with	with	ADP
cana-6094	147	8	provable	provable	ADJ
cana-6094	147	9	guarantees	guarantee	NOUN
cana-6094	147	10	for	for	ADP
cana-6094	147	11	capturing	capture	VERB
cana-6094	147	12	specific	specific	ADJ
cana-6094	147	13	mathematical	mathematical	ADJ
cana-6094	147	14	properties	property	NOUN
cana-6094	147	15	or	or	CCONJ
cana-6094	147	16	relationships	relationship	NOUN
cana-6094	147	17	110	110	NUM
cana-6094	147	18	.	.	PUNCT
cana-6094	148	1	mathematical	mathematical	ADJ
cana-6094	148	2	generalization	generalization	NOUN
cana-6094	148	3	theory	theory	NOUN
cana-6094	148	4	for	for	ADP
cana-6094	148	5	gnns	gnns	NOUN
cana-6094	148	6	would	would	AUX
cana-6094	148	7	provide	provide	VERB
cana-6094	148	8	formal	formal	ADJ
cana-6094	148	9	guarantees	guarantee	NOUN
cana-6094	148	10	about	about	ADP
cana-6094	148	11	their	their	PRON
cana-6094	148	12	performance	performance	NOUN
cana-6094	148	13	on	on	ADP
cana-6094	148	14	unseen	unseen	ADJ
cana-6094	148	15	mathematical	mathematical	ADJ
cana-6094	148	16	structures	structure	NOUN
cana-6094	148	17	,	,	PUNCT
cana-6094	148	18	particularly	particularly	ADV
cana-6094	148	19	those	those	PRON
cana-6094	148	20	that	that	PRON
cana-6094	148	21	differ	differ	VERB
cana-6094	148	22	systematically	systematically	ADV
cana-6094	148	23	from	from	ADP
cana-6094	148	24	training	train	VERB
cana-6094	148	25	examples	example	NOUN
cana-6094	148	26	.	.	PUNCT
cana-6094	149	1	such	such	ADJ
cana-6094	149	2	theory	theory	NOUN
cana-6094	149	3	would	would	AUX
cana-6094	149	4	increase	increase	VERB
cana-6094	149	5	confidence	confidence	NOUN
cana-6094	149	6	in	in	ADP
cana-6094	149	7	gnn	gnn	NOUN
cana-6094	149	8	predictions	prediction	NOUN
cana-6094	149	9	for	for	ADP
cana-6094	149	10	mathematical	mathematical	ADJ
cana-6094	149	11	discovery	discovery	NOUN
cana-6094	149	12	10	10	NUM
cana-6094	149	13	.	.	PUNCT
cana-6094	150	1	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	VERB
cana-6094	150	2	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	150	3	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	VERB
cana-6094	150	4	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	150	5	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	150	6	https://link.springer.com/article/10.1007/s11704-025-41426-w	https://link.springer.com/article/10.1007/s11704-025-41426-w	NOUN
cana-6094	150	7	https://rish-16.github.io/posts/gnn-math/	https://rish-16.github.io/posts/gnn-math/	NOUN
cana-6094	151	1	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	151	2	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	151	3	https://www.sciencedirect.com/science/article/pii/s2666651021000012	https://www.sciencedirect.com/science/article/pii/s2666651021000012	NOUN
cana-6094	151	4	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	151	5	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	151	6	communications	communication	NOUN
cana-6094	151	7	on	on	ADP
cana-6094	151	8	applied	apply	VERB
cana-6094	151	9	nonlinear	nonlinear	ADJ
cana-6094	151	10	analysis	analysis	NOUN
cana-6094	151	11	issn	issn	NOUN
cana-6094	151	12	:	:	PUNCT
cana-6094	151	13	1074	1074	NUM
cana-6094	151	14	-	-	PUNCT
cana-6094	151	15	133x	133x	NUM
cana-6094	151	16	vol	vol	NOUN
cana-6094	151	17	29	29	NUM
cana-6094	151	18	no	no	NOUN
cana-6094	151	19	.	.	NOUN
cana-6094	151	20	4	4	NUM
cana-6094	151	21	(	(	PUNCT
cana-6094	151	22	2022	2022	NUM
cana-6094	151	23	)	)	PUNCT
cana-6094	151	24	36	36	NUM
cana-6094	151	25	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	151	26	integration	integration	NOUN
cana-6094	151	27	with	with	ADP
cana-6094	151	28	formal	formal	ADJ
cana-6094	151	29	verification	verification	NOUN
cana-6094	151	30	could	could	AUX
cana-6094	151	31	provide	provide	VERB
cana-6094	151	32	frameworks	framework	NOUN
cana-6094	151	33	for	for	ADP
cana-6094	151	34	automatically	automatically	ADV
cana-6094	151	35	verifying	verify	VERB
cana-6094	151	36	gnn	gnn	NOUN
cana-6094	151	37	predictions	prediction	NOUN
cana-6094	151	38	against	against	ADP
cana-6094	151	39	mathematical	mathematical	ADJ
cana-6094	151	40	standards	standard	NOUN
cana-6094	151	41	.	.	PUNCT
cana-6094	152	1	these	these	DET
cana-6094	152	2	frameworks	framework	NOUN
cana-6094	152	3	would	would	AUX
cana-6094	152	4	combine	combine	VERB
cana-6094	152	5	statistical	statistical	ADJ
cana-6094	152	6	evidence	evidence	NOUN
cana-6094	152	7	from	from	ADP
cana-6094	152	8	gnns	gnns	NOUN
cana-6094	152	9	with	with	ADP
cana-6094	152	10	formal	formal	ADJ
cana-6094	152	11	proof	proof	NOUN
cana-6094	152	12	methods	method	NOUN
cana-6094	152	13	to	to	PART
cana-6094	152	14	create	create	VERB
cana-6094	152	15	hybrid	hybrid	ADJ
cana-6094	152	16	reasoning	reasoning	NOUN
cana-6094	152	17	systems	system	NOUN
cana-6094	152	18	with	with	ADP
cana-6094	152	19	both	both	DET
cana-6094	152	20	pattern	pattern	NOUN
cana-6094	152	21	recognition	recognition	NOUN
cana-6094	152	22	capabilities	capability	NOUN
cana-6094	152	23	and	and	CCONJ
cana-6094	152	24	mathematical	mathematical	ADJ
cana-6094	152	25	rigor	rigor	NOUN
cana-6094	152	26	910	910	NUM
cana-6094	152	27	.	.	PUNCT
cana-6094	153	1	7.conclusion	7.conclusion	NUM
cana-6094	153	2	graph	graph	NOUN
cana-6094	153	3	neural	neural	ADJ
cana-6094	153	4	networks	network	NOUN
cana-6094	153	5	represent	represent	VERB
cana-6094	153	6	a	a	DET
cana-6094	153	7	powerful	powerful	ADJ
cana-6094	153	8	paradigm	paradigm	NOUN
cana-6094	153	9	for	for	ADP
cana-6094	153	10	analyzing	analyze	VERB
cana-6094	153	11	complex	complex	ADJ
cana-6094	153	12	mathematical	mathematical	ADJ
cana-6094	153	13	structures	structure	NOUN
cana-6094	153	14	by	by	ADP
cana-6094	153	15	leveraging	leverage	VERB
cana-6094	153	16	their	their	PRON
cana-6094	153	17	inherent	inherent	ADJ
cana-6094	153	18	graph	graph	NOUN
cana-6094	153	19	-	-	PUNCT
cana-6094	153	20	like	like	ADJ
cana-6094	153	21	properties	property	NOUN
cana-6094	153	22	.	.	PUNCT
cana-6094	154	1	through	through	ADP
cana-6094	154	2	message	message	NOUN
cana-6094	154	3	passing	passing	NOUN
cana-6094	154	4	,	,	PUNCT
cana-6094	154	5	aggregation	aggregation	NOUN
cana-6094	154	6	,	,	PUNCT
cana-6094	154	7	and	and	CCONJ
cana-6094	154	8	update	update	NOUN
cana-6094	154	9	mechanisms	mechanism	NOUN
cana-6094	154	10	,	,	PUNCT
cana-6094	154	11	gnns	gnn	NOUN
cana-6094	154	12	capture	capture	VERB
cana-6094	154	13	the	the	DET
cana-6094	154	14	relational	relational	ADJ
cana-6094	154	15	patterns	pattern	NOUN
cana-6094	154	16	and	and	CCONJ
cana-6094	154	17	structural	structural	ADJ
cana-6094	154	18	dependencies	dependency	NOUN
cana-6094	154	19	that	that	PRON
cana-6094	154	20	characterize	characterize	VERB
cana-6094	154	21	mathematical	mathematical	ADJ
cana-6094	154	22	objects	object	NOUN
cana-6094	154	23	from	from	ADP
cana-6094	154	24	diverse	diverse	ADJ
cana-6094	154	25	domains	domain	NOUN
cana-6094	154	26	including	include	VERB
cana-6094	154	27	algebra	algebra	NOUN
cana-6094	154	28	,	,	PUNCT
cana-6094	154	29	geometry	geometry	NOUN
cana-6094	154	30	,	,	PUNCT
cana-6094	154	31	topology	topology	NOUN
cana-6094	154	32	,	,	PUNCT
cana-6094	154	33	and	and	CCONJ
cana-6094	154	34	combinatorics	combinatoric	NOUN
cana-6094	154	35	.	.	PUNCT
cana-6094	155	1	specialized	specialized	PROPN
cana-6094	155	2	gnn	gnn	PROPN
cana-6094	155	3	architectures	architecture	NOUN
cana-6094	155	4	including	include	VERB
cana-6094	155	5	spectral	spectral	ADJ
cana-6094	155	6	convolutional	convolutional	ADJ
cana-6094	155	7	networks	network	NOUN
cana-6094	155	8	,	,	PUNCT
cana-6094	155	9	attention	attention	NOUN
cana-6094	155	10	-	-	PUNCT
cana-6094	155	11	based	base	VERB
cana-6094	155	12	models	model	NOUN
cana-6094	155	13	,	,	PUNCT
cana-6094	155	14	and	and	CCONJ
cana-6094	155	15	geometrically	geometrically	ADV
cana-6094	155	16	equivariant	equivariant	ADJ
cana-6094	155	17	networks	network	NOUN
cana-6094	155	18	provide	provide	VERB
cana-6094	155	19	tailored	tailor	VERB
cana-6094	155	20	approaches	approach	NOUN
cana-6094	155	21	for	for	ADP
cana-6094	155	22	different	different	ADJ
cana-6094	155	23	mathematical	mathematical	ADJ
cana-6094	155	24	contexts	contexts	NOUN
cana-6094	155	25	,	,	PUNCT
cana-6094	155	26	preserving	preserve	VERB
cana-6094	155	27	relevant	relevant	ADJ
cana-6094	155	28	symmetries	symmetry	NOUN
cana-6094	155	29	and	and	CCONJ
cana-6094	155	30	focusing	focus	VERB
cana-6094	155	31	on	on	ADP
cana-6094	155	32	structurally	structurally	ADV
cana-6094	155	33	important	important	ADJ
cana-6094	155	34	relationships	relationship	NOUN
cana-6094	155	35	.	.	PUNCT
cana-6094	156	1	these	these	DET
cana-6094	156	2	architectures	architecture	NOUN
cana-6094	156	3	have	have	AUX
cana-6094	156	4	demonstrated	demonstrate	VERB
cana-6094	156	5	impressive	impressive	ADJ
cana-6094	156	6	capabilities	capability	NOUN
cana-6094	156	7	in	in	ADP
cana-6094	156	8	predicting	predict	VERB
cana-6094	156	9	mathematical	mathematical	ADJ
cana-6094	156	10	properties	property	NOUN
cana-6094	156	11	,	,	PUNCT
cana-6094	156	12	generating	generate	VERB
cana-6094	156	13	conjectures	conjecture	NOUN
cana-6094	156	14	,	,	PUNCT
cana-6094	156	15	and	and	CCONJ
cana-6094	156	16	even	even	ADV
cana-6094	156	17	contributing	contribute	VERB
cana-6094	156	18	to	to	AUX
cana-6094	156	19	theorem	theorem	ADJ
cana-6094	156	20	proving	proving	NOUN
cana-6094	156	21	when	when	SCONJ
cana-6094	156	22	combined	combine	VERB
cana-6094	156	23	with	with	ADP
cana-6094	156	24	traditional	traditional	ADJ
cana-6094	156	25	methods	method	NOUN
cana-6094	156	26	.	.	PUNCT
cana-6094	157	1	despite	despite	SCONJ
cana-6094	157	2	these	these	DET
cana-6094	157	3	successes	success	NOUN
cana-6094	157	4	,	,	PUNCT
cana-6094	157	5	significant	significant	ADJ
cana-6094	157	6	challenges	challenge	NOUN
cana-6094	157	7	remain	remain	VERB
cana-6094	157	8	in	in	ADP
cana-6094	157	9	providing	provide	VERB
cana-6094	157	10	theoretical	theoretical	ADJ
cana-6094	157	11	guarantees	guarantee	NOUN
cana-6094	157	12	,	,	PUNCT
cana-6094	157	13	addressing	address	VERB
cana-6094	157	14	data	datum	NOUN
cana-6094	157	15	scarcity	scarcity	NOUN
cana-6094	157	16	,	,	PUNCT
cana-6094	157	17	ensuring	ensure	VERB
cana-6094	157	18	mathematical	mathematical	ADJ
cana-6094	157	19	rigor	rigor	NOUN
cana-6094	157	20	,	,	PUNCT
cana-6094	157	21	and	and	CCONJ
cana-6094	157	22	achieving	achieve	VERB
cana-6094	157	23	interpretability	interpretability	NOUN
cana-6094	157	24	.	.	PUNCT
cana-6094	158	1	future	future	ADJ
cana-6094	158	2	research	research	NOUN
cana-6094	158	3	directions	direction	NOUN
cana-6094	158	4	point	point	VERB
cana-6094	158	5	toward	toward	ADP
cana-6094	158	6	integration	integration	NOUN
cana-6094	158	7	with	with	ADP
cana-6094	158	8	symbolic	symbolic	ADJ
cana-6094	158	9	ai	ai	PROPN
cana-6094	158	10	,	,	PUNCT
cana-6094	158	11	advanced	advanced	ADJ
cana-6094	158	12	geometrical	geometrical	ADJ
cana-6094	158	13	learning	learning	NOUN
cana-6094	158	14	,	,	PUNCT
cana-6094	158	15	cross	cross	ADJ
cana-6094	158	16	-	-	ADJ
cana-6094	158	17	domain	domain	ADJ
cana-6094	158	18	applications	application	NOUN
cana-6094	158	19	,	,	PUNCT
cana-6094	158	20	and	and	CCONJ
cana-6094	158	21	theoretical	theoretical	ADJ
cana-6094	158	22	advances	advance	NOUN
cana-6094	158	23	that	that	PRON
cana-6094	158	24	will	will	AUX
cana-6094	158	25	further	far	ADV
cana-6094	158	26	strengthen	strengthen	VERB
cana-6094	158	27	the	the	DET
cana-6094	158	28	role	role	NOUN
cana-6094	158	29	of	of	ADP
cana-6094	158	30	gnns	gnns	NOUN
cana-6094	158	31	in	in	ADP
cana-6094	158	32	mathematical	mathematical	ADJ
cana-6094	158	33	analysis	analysis	NOUN
cana-6094	158	34	.	.	PUNCT
cana-6094	159	1	as	as	SCONJ
cana-6094	159	2	gnn	gnn	PROPN
cana-6094	159	3	methodologies	methodology	NOUN
cana-6094	159	4	continue	continue	VERB
cana-6094	159	5	to	to	PART
cana-6094	159	6	evolve	evolve	VERB
cana-6094	159	7	and	and	CCONJ
cana-6094	159	8	integrate	integrate	VERB
cana-6094	159	9	with	with	ADP
cana-6094	159	10	traditional	traditional	ADJ
cana-6094	159	11	mathematical	mathematical	ADJ
cana-6094	159	12	approaches	approach	NOUN
cana-6094	159	13	,	,	PUNCT
cana-6094	159	14	they	they	PRON
cana-6094	159	15	hold	hold	VERB
cana-6094	159	16	the	the	DET
cana-6094	159	17	potential	potential	NOUN
cana-6094	159	18	to	to	PART
cana-6094	159	19	transform	transform	VERB
cana-6094	159	20	how	how	SCONJ
cana-6094	159	21	we	we	PRON
cana-6094	159	22	explore	explore	VERB
cana-6094	159	23	and	and	CCONJ
cana-6094	159	24	understand	understand	VERB
cana-6094	159	25	complex	complex	ADJ
cana-6094	159	26	mathematical	mathematical	ADJ
cana-6094	159	27	structures	structure	NOUN
cana-6094	159	28	,	,	PUNCT
cana-6094	159	29	accelerating	accelerate	VERB
cana-6094	159	30	mathematical	mathematical	ADJ
cana-6094	159	31	discovery	discovery	NOUN
cana-6094	159	32	through	through	ADP
cana-6094	159	33	data	data	NOUN
cana-6094	159	34	-	-	PUNCT
cana-6094	159	35	driven	drive	VERB
cana-6094	159	36	pattern	pattern	NOUN
cana-6094	159	37	recognition	recognition	NOUN
cana-6094	159	38	complemented	complement	VERB
cana-6094	159	39	by	by	ADP
cana-6094	159	40	formal	formal	ADJ
cana-6094	159	41	reasoning	reasoning	NOUN
cana-6094	159	42	.	.	PUNCT
cana-6094	160	1	this	this	DET
cana-6094	160	2	synergy	synergy	NOUN
cana-6094	160	3	between	between	ADP
cana-6094	160	4	statistical	statistical	ADJ
cana-6094	160	5	learning	learning	NOUN
cana-6094	160	6	and	and	CCONJ
cana-6094	160	7	mathematical	mathematical	ADJ
cana-6094	160	8	rigor	rigor	NOUN
cana-6094	160	9	represents	represent	VERB
cana-6094	160	10	an	an	DET
cana-6094	160	11	exciting	exciting	ADJ
cana-6094	160	12	frontier	frontier	NOUN
cana-6094	160	13	in	in	ADP
cana-6094	160	14	both	both	PRON
cana-6094	160	15	machine	machine	NOUN
cana-6094	160	16	learning	learning	NOUN
cana-6094	160	17	and	and	CCONJ
cana-6094	160	18	mathematical	mathematical	ADJ
cana-6094	160	19	sciences	science	NOUN
cana-6094	160	20	.	.	PUNCT
cana-6094	161	1	references	reference	NOUN
cana-6094	161	2	1	1	NUM
cana-6094	161	3	.	.	PUNCT
cana-6094	161	4	scarselli	scarselli	PROPN
cana-6094	161	5	,	,	PUNCT
cana-6094	161	6	f.	f.	PROPN
cana-6094	161	7	,	,	PUNCT
cana-6094	161	8	gori	gori	PROPN
cana-6094	161	9	,	,	PUNCT
cana-6094	161	10	m.	m.	NOUN
cana-6094	161	11	,	,	PUNCT
cana-6094	161	12	tsoi	tsoi	PROPN
cana-6094	161	13	,	,	PUNCT
cana-6094	161	14	a.	a.	PROPN
cana-6094	161	15	c.	c.	PROPN
cana-6094	161	16	,	,	PUNCT
cana-6094	161	17	hagenbuchner	hagenbuchner	ADJ
cana-6094	161	18	,	,	PUNCT
cana-6094	161	19	m.	m.	NOUN
cana-6094	161	20	,	,	PUNCT
cana-6094	161	21	&	&	CCONJ
cana-6094	161	22	monfardini	monfardini	PROPN
cana-6094	161	23	,	,	PUNCT
cana-6094	161	24	g.	g.	PROPN
cana-6094	161	25	(	(	PUNCT
cana-6094	161	26	2009	2009	NUM
cana-6094	161	27	)	)	PUNCT
cana-6094	161	28	.	.	PUNCT
cana-6094	162	1	the	the	DET
cana-6094	162	2	graph	graph	NOUN
cana-6094	162	3	neural	neural	ADJ
cana-6094	162	4	network	network	NOUN
cana-6094	162	5	model	model	NOUN
cana-6094	162	6	.	.	PUNCT
cana-6094	163	1	ieee	ieee	NOUN
cana-6094	163	2	transactions	transaction	NOUN
cana-6094	163	3	on	on	ADP
cana-6094	163	4	neural	neural	ADJ
cana-6094	163	5	networks	network	NOUN
cana-6094	163	6	,	,	PUNCT
cana-6094	163	7	*	*	NOUN
cana-6094	163	8	20*(1	20*(1	NUM
cana-6094	163	9	)	)	PUNCT
cana-6094	163	10	,	,	PUNCT
cana-6094	163	11	61–80	61–80	NOUN
cana-6094	163	12	.	.	NOUN
cana-6094	164	1	2	2	NUM
cana-6094	164	2	.	.	X
cana-6094	164	3	bruna	bruna	PROPN
cana-6094	164	4	,	,	PUNCT
cana-6094	164	5	j.	j.	PROPN
cana-6094	164	6	,	,	PUNCT
cana-6094	164	7	zaremba	zaremba	PROPN
cana-6094	164	8	,	,	PUNCT
cana-6094	164	9	w.	w.	PROPN
cana-6094	164	10	,	,	PUNCT
cana-6094	164	11	szlam	szlam	PROPN
cana-6094	164	12	,	,	PUNCT
cana-6094	164	13	a.	a.	PROPN
cana-6094	164	14	,	,	PUNCT
cana-6094	164	15	&	&	CCONJ
cana-6094	164	16	lecun	lecun	PROPN
cana-6094	164	17	,	,	PUNCT
cana-6094	164	18	y.	y.	PROPN
cana-6094	164	19	(	(	PUNCT
cana-6094	164	20	2014	2014	NUM
cana-6094	164	21	)	)	PUNCT
cana-6094	164	22	.	.	PUNCT
cana-6094	165	1	spectral	spectral	ADJ
cana-6094	165	2	networks	network	NOUN
cana-6094	165	3	and	and	CCONJ
cana-6094	165	4	locally	locally	ADV
cana-6094	165	5	connected	connect	VERB
cana-6094	165	6	networks	network	NOUN
cana-6094	165	7	on	on	ADP
cana-6094	165	8	graphs	graph	NOUN
cana-6094	165	9	.	.	PUNCT
cana-6094	166	1	international	international	ADJ
cana-6094	166	2	conference	conference	NOUN
cana-6094	166	3	on	on	ADP
cana-6094	166	4	learning	learn	VERB
cana-6094	166	5	representations	representation	NOUN
cana-6094	166	6	(	(	PUNCT
cana-6094	166	7	iclr	iclr	NOUN
cana-6094	166	8	)	)	PUNCT
cana-6094	166	9	.	.	PUNCT
cana-6094	167	1	3	3	X
cana-6094	167	2	.	.	X
cana-6094	167	3	kipf	kipf	PROPN
cana-6094	167	4	,	,	PUNCT
cana-6094	167	5	t.	t.	PROPN
cana-6094	167	6	n.	n.	PROPN
cana-6094	167	7	,	,	PUNCT
cana-6094	167	8	&	&	CCONJ
cana-6094	167	9	welling	well	VERB
cana-6094	167	10	,	,	PUNCT
cana-6094	167	11	m.	m.	NOUN
cana-6094	167	12	(	(	PUNCT
cana-6094	167	13	2017	2017	NUM
cana-6094	167	14	)	)	PUNCT
cana-6094	167	15	.	.	PUNCT
cana-6094	168	1	semi	semi	ADJ
cana-6094	168	2	-	-	ADJ
cana-6094	168	3	supervised	supervised	ADJ
cana-6094	168	4	classification	classification	NOUN
cana-6094	168	5	with	with	ADP
cana-6094	168	6	graph	graph	NOUN
cana-6094	168	7	convolutional	convolutional	ADJ
cana-6094	168	8	networks	network	NOUN
cana-6094	168	9	.	.	PUNCT
cana-6094	169	1	international	international	ADJ
cana-6094	169	2	conference	conference	NOUN
cana-6094	169	3	on	on	ADP
cana-6094	169	4	learning	learn	VERB
cana-6094	169	5	representations	representation	NOUN
cana-6094	169	6	(	(	PUNCT
cana-6094	169	7	iclr	iclr	NOUN
cana-6094	169	8	)	)	PUNCT
cana-6094	169	9	.	.	PUNCT
cana-6094	170	1	4	4	X
cana-6094	170	2	.	.	X
cana-6094	170	3	hamilton	hamilton	PROPN
cana-6094	170	4	,	,	PUNCT
cana-6094	170	5	w.	w.	PROPN
cana-6094	170	6	l.	l.	PROPN
cana-6094	170	7	,	,	PUNCT
cana-6094	170	8	ying	ying	PROPN
cana-6094	170	9	,	,	PUNCT
cana-6094	170	10	r.	r.	PROPN
cana-6094	170	11	,	,	PUNCT
cana-6094	170	12	&	&	CCONJ
cana-6094	170	13	leskovec	leskovec	PROPN
cana-6094	170	14	,	,	PUNCT
cana-6094	170	15	j.	j.	PROPN
cana-6094	170	16	(	(	PUNCT
cana-6094	170	17	2017	2017	NUM
cana-6094	170	18	)	)	PUNCT
cana-6094	170	19	.	.	PUNCT
cana-6094	171	1	inductive	inductive	ADJ
cana-6094	171	2	representation	representation	NOUN
cana-6094	171	3	learning	learn	VERB
cana-6094	171	4	on	on	ADP
cana-6094	171	5	large	large	ADJ
cana-6094	171	6	graphs	graph	NOUN
cana-6094	171	7	.	.	PUNCT
cana-6094	172	1	advances	advance	NOUN
cana-6094	172	2	in	in	ADP
cana-6094	172	3	neural	neural	ADJ
cana-6094	172	4	information	information	NOUN
cana-6094	172	5	processing	processing	NOUN
cana-6094	172	6	systems	system	NOUN
cana-6094	172	7	(	(	PUNCT
cana-6094	172	8	neurips	neurip	NOUN
cana-6094	172	9	)	)	PUNCT
cana-6094	172	10	.	.	PUNCT
cana-6094	173	1	5	5	X
cana-6094	173	2	.	.	X
cana-6094	173	3	velickovic	velickovic	PROPN
cana-6094	173	4	,	,	PUNCT
cana-6094	173	5	p.	p.	NOUN
cana-6094	173	6	,	,	PUNCT
cana-6094	173	7	cucurull	cucurull	NOUN
cana-6094	173	8	,	,	PUNCT
cana-6094	173	9	g.	g.	PROPN
cana-6094	173	10	,	,	PUNCT
cana-6094	173	11	casanova	casanova	PROPN
cana-6094	173	12	,	,	PUNCT
cana-6094	173	13	a.	a.	PROPN
cana-6094	173	14	,	,	PUNCT
cana-6094	173	15	romero	romero	PROPN
cana-6094	173	16	,	,	PUNCT
cana-6094	173	17	a.	a.	PROPN
cana-6094	173	18	,	,	PUNCT
cana-6094	173	19	lio	lio	PROPN
cana-6094	173	20	,	,	PUNCT
cana-6094	173	21	p.	p.	PROPN
cana-6094	173	22	,	,	PUNCT
cana-6094	173	23	&	&	CCONJ
cana-6094	173	24	bengio	bengio	PROPN
cana-6094	173	25	,	,	PUNCT
cana-6094	173	26	y.	y.	PROPN
cana-6094	173	27	(	(	PUNCT
cana-6094	173	28	2018	2018	NUM
cana-6094	173	29	)	)	PUNCT
cana-6094	173	30	.	.	PUNCT
cana-6094	174	1	graph	graph	NOUN
cana-6094	174	2	attention	attention	NOUN
cana-6094	174	3	networks	network	NOUN
cana-6094	174	4	.	.	PUNCT
cana-6094	175	1	international	international	ADJ
cana-6094	175	2	conference	conference	NOUN
cana-6094	175	3	on	on	ADP
cana-6094	175	4	learning	learn	VERB
cana-6094	175	5	representations	representation	NOUN
cana-6094	175	6	(	(	PUNCT
cana-6094	175	7	iclr	iclr	NOUN
cana-6094	175	8	)	)	PUNCT
cana-6094	175	9	.	.	PUNCT
cana-6094	176	1	6	6	X
cana-6094	176	2	.	.	X
cana-6094	176	3	defferrard	defferrard	NOUN
cana-6094	176	4	,	,	PUNCT
cana-6094	176	5	m.	m.	NOUN
cana-6094	176	6	,	,	PUNCT
cana-6094	176	7	bresson	bresson	PROPN
cana-6094	176	8	,	,	PUNCT
cana-6094	176	9	x.	x.	PROPN
cana-6094	176	10	,	,	PUNCT
cana-6094	176	11	&	&	CCONJ
cana-6094	176	12	vandergheynst	vandergheynst	PROPN
cana-6094	176	13	,	,	PUNCT
cana-6094	176	14	p.	p.	NOUN
cana-6094	176	15	(	(	PUNCT
cana-6094	176	16	2016	2016	NUM
cana-6094	176	17	)	)	PUNCT
cana-6094	176	18	.	.	PUNCT
cana-6094	177	1	convolutional	convolutional	ADJ
cana-6094	177	2	neural	neural	ADJ
cana-6094	177	3	networks	network	NOUN
cana-6094	177	4	on	on	ADP
cana-6094	177	5	graphs	graph	NOUN
cana-6094	177	6	with	with	ADP
cana-6094	177	7	fast	fast	ADJ
cana-6094	177	8	localized	localize	VERB
cana-6094	177	9	spectral	spectral	ADJ
cana-6094	177	10	filtering	filtering	NOUN
cana-6094	177	11	.	.	PUNCT
cana-6094	178	1	advances	advance	NOUN
cana-6094	178	2	in	in	ADP
cana-6094	178	3	neural	neural	ADJ
cana-6094	178	4	information	information	NOUN
cana-6094	178	5	processing	processing	NOUN
cana-6094	178	6	systems	system	NOUN
cana-6094	178	7	(	(	PUNCT
cana-6094	178	8	neurips	neurip	NOUN
cana-6094	178	9	)	)	PUNCT
cana-6094	178	10	.	.	PUNCT
cana-6094	179	1	7	7	X
cana-6094	179	2	.	.	X
cana-6094	179	3	battaglia	battaglia	PROPN
cana-6094	179	4	,	,	PUNCT
cana-6094	179	5	p.	p.	PROPN
cana-6094	179	6	w.	w.	PROPN
cana-6094	179	7	,	,	PUNCT
cana-6094	179	8	hamrick	hamrick	PROPN
cana-6094	179	9	,	,	PUNCT
cana-6094	179	10	j.	j.	PROPN
cana-6094	179	11	b.	b.	PROPN
cana-6094	179	12	,	,	PUNCT
cana-6094	179	13	bapst	bapst	NOUN
cana-6094	179	14	,	,	PUNCT
cana-6094	179	15	v.	v.	PROPN
cana-6094	179	16	,	,	PUNCT
cana-6094	179	17	sanchez	sanchez	PROPN
cana-6094	179	18	-	-	PUNCT
cana-6094	179	19	gonzalez	gonzalez	PROPN
cana-6094	179	20	,	,	PUNCT
cana-6094	179	21	a.	a.	PROPN
cana-6094	179	22	,	,	PUNCT
cana-6094	179	23	zambaldi	zambaldi	PROPN
cana-6094	179	24	,	,	PUNCT
cana-6094	179	25	v.	v.	PROPN
cana-6094	179	26	,	,	PUNCT
cana-6094	179	27	malinowski	malinowski	NOUN
cana-6094	179	28	,	,	PUNCT
cana-6094	179	29	m.	m.	NOUN
cana-6094	179	30	,	,	PUNCT
cana-6094	179	31	...	...	PUNCT
cana-6094	179	32	&	&	CCONJ
cana-6094	179	33	pascanu	pascanu	PROPN
cana-6094	179	34	,	,	PUNCT
cana-6094	179	35	r.	r.	PROPN
cana-6094	179	36	(	(	PUNCT
cana-6094	179	37	2018	2018	NUM
cana-6094	179	38	)	)	PUNCT
cana-6094	179	39	.	.	PUNCT
cana-6094	180	1	relational	relational	ADJ
cana-6094	180	2	inductive	inductive	PROPN
cana-6094	180	3	biases	bias	NOUN
cana-6094	180	4	,	,	PUNCT
cana-6094	180	5	deep	deep	ADJ
cana-6094	180	6	learning	learning	NOUN
cana-6094	180	7	,	,	PUNCT
cana-6094	180	8	and	and	CCONJ
cana-6094	180	9	graph	graph	NOUN
cana-6094	180	10	networks	network	NOUN
cana-6094	180	11	.	.	PUNCT
cana-6094	181	1	arxiv	arxiv	PROPN
cana-6094	181	2	preprint	preprint	PROPN
cana-6094	181	3	arxiv:1806.01261	arxiv:1806.01261	PROPN
cana-6094	181	4	.	.	PUNCT
cana-6094	182	1	8	8	NUM
cana-6094	182	2	.	.	X
cana-6094	182	3	bronstein	bronstein	PROPN
cana-6094	182	4	,	,	PUNCT
cana-6094	182	5	m.	m.	NOUN
cana-6094	182	6	m.	m.	PROPN
cana-6094	182	7	,	,	PUNCT
cana-6094	182	8	bruna	bruna	PROPN
cana-6094	182	9	,	,	PUNCT
cana-6094	182	10	j.	j.	PROPN
cana-6094	182	11	,	,	PUNCT
cana-6094	182	12	cohen	cohen	PROPN
cana-6094	182	13	,	,	PUNCT
cana-6094	182	14	t.	t.	PROPN
cana-6094	182	15	,	,	PUNCT
cana-6094	182	16	&	&	CCONJ
cana-6094	182	17	veličković	veličković	PROPN
cana-6094	182	18	,	,	PUNCT
cana-6094	182	19	p.	p.	NOUN
cana-6094	182	20	(	(	PUNCT
cana-6094	182	21	2021	2021	NUM
cana-6094	182	22	)	)	PUNCT
cana-6094	182	23	.	.	PUNCT
cana-6094	183	1	geometric	geometric	ADJ
cana-6094	183	2	deep	deep	ADJ
cana-6094	183	3	learning	learning	NOUN
cana-6094	183	4	:	:	PUNCT
cana-6094	183	5	grids	grid	NOUN
cana-6094	183	6	,	,	PUNCT
cana-6094	183	7	groups	group	NOUN
cana-6094	183	8	,	,	PUNCT
cana-6094	183	9	graphs	graph	NOUN
cana-6094	183	10	,	,	PUNCT
cana-6094	183	11	geodesics	geodesic	NOUN
cana-6094	183	12	,	,	PUNCT
cana-6094	183	13	and	and	CCONJ
cana-6094	183	14	gauges	gauge	NOUN
cana-6094	183	15	.	.	PUNCT
cana-6094	184	1	arxiv	arxiv	PROPN
cana-6094	184	2	preprint	preprint	PROPN
cana-6094	184	3	arxiv:2104.13478	arxiv:2104.13478	PROPN
cana-6094	184	4	.	.	PUNCT
cana-6094	185	1	9	9	X
cana-6094	185	2	.	.	X
cana-6094	186	1	gilmer	gilmer	PROPN
cana-6094	186	2	,	,	PUNCT
cana-6094	186	3	j.	j.	PROPN
cana-6094	186	4	,	,	PUNCT
cana-6094	186	5	schoenholz	schoenholz	PROPN
cana-6094	186	6	,	,	PUNCT
cana-6094	186	7	s.	s.	PROPN
cana-6094	186	8	s.	s.	PROPN
cana-6094	186	9	,	,	PUNCT
cana-6094	186	10	riley	riley	PROPN
cana-6094	186	11	,	,	PUNCT
cana-6094	186	12	p.	p.	PROPN
cana-6094	186	13	f.	f.	PROPN
cana-6094	186	14	,	,	PUNCT
cana-6094	186	15	vinyals	vinyal	NOUN
cana-6094	186	16	,	,	PUNCT
cana-6094	186	17	o.	o.	PROPN
cana-6094	186	18	,	,	PUNCT
cana-6094	186	19	&	&	CCONJ
cana-6094	186	20	dahl	dahl	PROPN
cana-6094	186	21	,	,	PUNCT
cana-6094	186	22	g.	g.	PROPN
cana-6094	186	23	e.	e.	PROPN
cana-6094	186	24	(	(	PUNCT
cana-6094	186	25	2017	2017	NUM
cana-6094	186	26	)	)	PUNCT
cana-6094	186	27	.	.	PUNCT
cana-6094	187	1	neural	neural	ADJ
cana-6094	187	2	message	message	NOUN
cana-6094	187	3	passing	pass	VERB
cana-6094	187	4	for	for	ADP
cana-6094	187	5	quantum	quantum	ADJ
cana-6094	187	6	chemistry	chemistry	NOUN
cana-6094	187	7	.	.	PUNCT
cana-6094	188	1	international	international	ADJ
cana-6094	188	2	conference	conference	NOUN
cana-6094	188	3	on	on	ADP
cana-6094	188	4	machine	machine	NOUN
cana-6094	188	5	learning	learning	NOUN
cana-6094	188	6	(	(	PUNCT
cana-6094	188	7	icml	icml	PROPN
cana-6094	188	8	)	)	PUNCT
cana-6094	188	9	.	.	PUNCT
cana-6094	189	1	10	10	NUM
cana-6094	189	2	.	.	X
cana-6094	189	3	cohen	cohen	PROPN
cana-6094	189	4	,	,	PUNCT
cana-6094	189	5	t.	t.	PROPN
cana-6094	189	6	s.	s.	PROPN
cana-6094	189	7	,	,	PUNCT
cana-6094	189	8	&	&	CCONJ
cana-6094	189	9	welling	well	VERB
cana-6094	189	10	,	,	PUNCT
cana-6094	189	11	m.	m.	NOUN
cana-6094	189	12	(	(	PUNCT
cana-6094	189	13	2016	2016	NUM
cana-6094	189	14	)	)	PUNCT
cana-6094	189	15	.	.	PUNCT
cana-6094	190	1	group	group	PROPN
cana-6094	190	2	equivariant	equivariant	ADJ
cana-6094	190	3	convolutional	convolutional	ADJ
cana-6094	190	4	networks	network	NOUN
cana-6094	190	5	.	.	PUNCT
cana-6094	191	1	international	international	ADJ
cana-6094	191	2	conference	conference	NOUN
cana-6094	191	3	on	on	ADP
cana-6094	191	4	machine	machine	NOUN
cana-6094	191	5	learning	learning	NOUN
cana-6094	191	6	(	(	PUNCT
cana-6094	191	7	icml	icml	PROPN
cana-6094	191	8	)	)	PUNCT
cana-6094	191	9	.	.	PUNCT
cana-6094	192	1	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	https://jonathan-hui.medium.com/applications-of-graph-neural-networks-gnn-d487fd5ed17d	NOUN
cana-6094	192	2	https://www.nature.com/articles/s41598-025-12840-8	https://www.nature.com/articles/s41598-025-12840-8	VERB
cana-6094	192	3	communications	communication	NOUN
cana-6094	192	4	on	on	ADP
cana-6094	192	5	applied	apply	VERB
cana-6094	192	6	nonlinear	nonlinear	ADJ
cana-6094	192	7	analysis	analysis	NOUN
cana-6094	192	8	issn	issn	NOUN
cana-6094	192	9	:	:	PUNCT
cana-6094	192	10	1074	1074	NUM
cana-6094	192	11	-	-	PUNCT
cana-6094	192	12	133x	133x	NUM
cana-6094	192	13	vol	vol	NOUN
cana-6094	192	14	29	29	NUM
cana-6094	192	15	no	no	NOUN
cana-6094	192	16	.	.	NOUN
cana-6094	192	17	4	4	NUM
cana-6094	192	18	(	(	PUNCT
cana-6094	192	19	2022	2022	NUM
cana-6094	192	20	)	)	PUNCT
cana-6094	192	21	37	37	NUM
cana-6094	192	22	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	192	23	11	11	NUM
cana-6094	192	24	.	.	PUNCT
cana-6094	193	1	satorras	satorra	NOUN
cana-6094	193	2	,	,	PUNCT
cana-6094	193	3	v.	v.	PROPN
cana-6094	193	4	g.	g.	PROPN
cana-6094	193	5	,	,	PUNCT
cana-6094	193	6	hoogeboom	hoogeboom	PROPN
cana-6094	193	7	,	,	PUNCT
cana-6094	193	8	e.	e.	PROPN
cana-6094	193	9	,	,	PUNCT
cana-6094	193	10	&	&	CCONJ
cana-6094	193	11	welling	well	VERB
cana-6094	193	12	,	,	PUNCT
cana-6094	193	13	m.	m.	NOUN
cana-6094	193	14	(	(	PUNCT
cana-6094	193	15	2021	2021	NUM
cana-6094	193	16	)	)	PUNCT
cana-6094	193	17	.	.	PUNCT
cana-6094	194	1	e(n	e(n	ADJ
cana-6094	194	2	)	)	PUNCT
cana-6094	194	3	equivariant	equivariant	PROPN
cana-6094	194	4	graph	graph	NOUN
cana-6094	194	5	neural	neural	ADJ
cana-6094	194	6	networks	network	NOUN
cana-6094	194	7	.	.	PUNCT
cana-6094	195	1	international	international	ADJ
cana-6094	195	2	conference	conference	NOUN
cana-6094	195	3	on	on	ADP
cana-6094	195	4	machine	machine	NOUN
cana-6094	195	5	learning	learning	NOUN
cana-6094	195	6	(	(	PUNCT
cana-6094	195	7	icml	icml	PROPN
cana-6094	195	8	)	)	PUNCT
cana-6094	195	9	.	.	PUNCT
cana-6094	196	1	12	12	NUM
cana-6094	196	2	.	.	PUNCT
cana-6094	197	1	thomas	thomas	PROPN
cana-6094	197	2	,	,	PUNCT
cana-6094	197	3	n.	n.	NOUN
cana-6094	197	4	,	,	PUNCT
cana-6094	197	5	smidt	smidt	ADJ
cana-6094	197	6	,	,	PUNCT
cana-6094	197	7	t.	t.	PROPN
cana-6094	197	8	,	,	PUNCT
cana-6094	197	9	kearnes	kearne	NOUN
cana-6094	197	10	,	,	PUNCT
cana-6094	197	11	s.	s.	PROPN
cana-6094	197	12	,	,	PUNCT
cana-6094	197	13	yang	yang	PROPN
cana-6094	197	14	,	,	PUNCT
cana-6094	197	15	l.	l.	PROPN
cana-6094	197	16	,	,	PUNCT
cana-6094	197	17	li	li	PROPN
cana-6094	197	18	,	,	PUNCT
cana-6094	197	19	l.	l.	PROPN
cana-6094	197	20	,	,	PUNCT
cana-6094	197	21	kohlhoff	kohlhoff	PROPN
cana-6094	197	22	,	,	PUNCT
cana-6094	197	23	k.	k.	PROPN
cana-6094	197	24	,	,	PUNCT
cana-6094	197	25	&	&	CCONJ
cana-6094	197	26	riley	riley	PROPN
cana-6094	197	27	,	,	PUNCT
cana-6094	197	28	p.	p.	NOUN
cana-6094	197	29	(	(	PUNCT
cana-6094	197	30	2018	2018	NUM
cana-6094	197	31	)	)	PUNCT
cana-6094	197	32	.	.	PUNCT
cana-6094	198	1	tensor	tensor	NOUN
cana-6094	198	2	field	field	NOUN
cana-6094	198	3	networks	network	NOUN
cana-6094	198	4	:	:	PUNCT
cana-6094	198	5	rotation	rotation	NOUN
cana-6094	198	6	-	-	PUNCT
cana-6094	198	7	and	and	CCONJ
cana-6094	198	8	translation	translation	NOUN
cana-6094	198	9	-	-	PUNCT
cana-6094	198	10	equivariant	equivariant	ADJ
cana-6094	198	11	neural	neural	ADJ
cana-6094	198	12	networks	network	NOUN
cana-6094	198	13	for	for	ADP
cana-6094	198	14	3d	3d	NUM
cana-6094	198	15	point	point	NOUN
cana-6094	198	16	clouds	cloud	NOUN
cana-6094	198	17	.	.	PUNCT
cana-6094	199	1	arxiv	arxiv	PROPN
cana-6094	199	2	preprint	preprint	VERB
cana-6094	199	3	arxiv:1802.08219	arxiv:1802.08219	ADV
cana-6094	199	4	.	.	PUNCT
cana-6094	200	1	13	13	NUM
cana-6094	200	2	.	.	X
cana-6094	201	1	finzi	finzi	PROPN
cana-6094	201	2	,	,	PUNCT
cana-6094	201	3	m.	m.	NOUN
cana-6094	201	4	,	,	PUNCT
cana-6094	201	5	stanton	stanton	PROPN
cana-6094	201	6	,	,	PUNCT
cana-6094	201	7	s.	s.	PROPN
cana-6094	201	8	,	,	PUNCT
cana-6094	201	9	izmailov	izmailov	PROPN
cana-6094	201	10	,	,	PUNCT
cana-6094	201	11	p.	p.	PROPN
cana-6094	201	12	,	,	PUNCT
cana-6094	201	13	&	&	CCONJ
cana-6094	201	14	wilson	wilson	PROPN
cana-6094	201	15	,	,	PUNCT
cana-6094	201	16	a.	a.	PROPN
cana-6094	201	17	g.	g.	PROPN
cana-6094	201	18	(	(	PUNCT
cana-6094	201	19	2021	2021	NUM
cana-6094	201	20	)	)	PUNCT
cana-6094	201	21	.	.	PUNCT
cana-6094	202	1	generalizing	generalize	VERB
cana-6094	202	2	convolutional	convolutional	ADJ
cana-6094	202	3	neural	neural	ADJ
cana-6094	202	4	networks	network	NOUN
cana-6094	202	5	for	for	ADP
cana-6094	202	6	equivariance	equivariance	NOUN
cana-6094	202	7	to	to	PART
cana-6094	202	8	lie	lie	VERB
cana-6094	202	9	groups	group	NOUN
cana-6094	202	10	on	on	ADP
cana-6094	202	11	arbitrary	arbitrary	ADJ
cana-6094	202	12	continuous	continuous	ADJ
cana-6094	202	13	data	datum	NOUN
cana-6094	202	14	.	.	PUNCT
cana-6094	203	1	international	international	ADJ
cana-6094	203	2	conference	conference	NOUN
cana-6094	203	3	on	on	ADP
cana-6094	203	4	machine	machine	NOUN
cana-6094	203	5	learning	learning	NOUN
cana-6094	203	6	(	(	PUNCT
cana-6094	203	7	icml	icml	PROPN
cana-6094	203	8	)	)	PUNCT
cana-6094	203	9	.	.	PUNCT
cana-6094	204	1	14	14	NUM
cana-6094	204	2	.	.	PUNCT
cana-6094	205	1	xu	xu	PROPN
cana-6094	205	2	,	,	PUNCT
cana-6094	205	3	k.	k.	PROPN
cana-6094	205	4	,	,	PUNCT
cana-6094	205	5	hu	hu	PROPN
cana-6094	205	6	,	,	PUNCT
cana-6094	205	7	w.	w.	PROPN
cana-6094	205	8	,	,	PUNCT
cana-6094	205	9	leskovec	leskovec	PROPN
cana-6094	205	10	,	,	PUNCT
cana-6094	205	11	j.	j.	PROPN
cana-6094	205	12	,	,	PUNCT
cana-6094	205	13	&	&	CCONJ
cana-6094	205	14	jegelka	jegelka	PROPN
cana-6094	205	15	,	,	PUNCT
cana-6094	205	16	s.	s.	PROPN
cana-6094	205	17	(	(	PUNCT
cana-6094	205	18	2019	2019	NUM
cana-6094	205	19	)	)	PUNCT
cana-6094	205	20	.	.	PUNCT
cana-6094	206	1	how	how	SCONJ
cana-6094	206	2	powerful	powerful	ADJ
cana-6094	206	3	are	be	AUX
cana-6094	206	4	graph	graph	VERB
cana-6094	206	5	neural	neural	ADJ
cana-6094	206	6	networks	network	NOUN
cana-6094	206	7	?	?	PUNCT
cana-6094	207	1	international	international	ADJ
cana-6094	207	2	conference	conference	NOUN
cana-6094	207	3	on	on	ADP
cana-6094	207	4	learning	learn	VERB
cana-6094	207	5	representations	representation	NOUN
cana-6094	207	6	(	(	PUNCT
cana-6094	207	7	iclr	iclr	NOUN
cana-6094	207	8	)	)	PUNCT
cana-6094	207	9	.	.	PUNCT
cana-6094	208	1	15	15	X
cana-6094	208	2	.	.	X
cana-6094	209	1	morris	morris	PROPN
cana-6094	209	2	,	,	PUNCT
cana-6094	209	3	c.	c.	PROPN
cana-6094	209	4	,	,	PUNCT
cana-6094	209	5	ritzert	ritzert	ADJ
cana-6094	209	6	,	,	PUNCT
cana-6094	209	7	m.	m.	NOUN
cana-6094	209	8	,	,	PUNCT
cana-6094	209	9	fey	fey	PROPN
cana-6094	209	10	,	,	PUNCT
cana-6094	209	11	m.	m.	NOUN
cana-6094	209	12	,	,	PUNCT
cana-6094	209	13	hamilton	hamilton	PROPN
cana-6094	209	14	,	,	PUNCT
cana-6094	209	15	w.	w.	PROPN
cana-6094	209	16	l.	l.	PROPN
cana-6094	209	17	,	,	PUNCT
cana-6094	209	18	lenssen	lenssen	PROPN
cana-6094	209	19	,	,	PUNCT
cana-6094	209	20	j.	j.	PROPN
cana-6094	209	21	e.	e.	PROPN
cana-6094	209	22	,	,	PUNCT
cana-6094	209	23	rattan	rattan	PROPN
cana-6094	209	24	,	,	PUNCT
cana-6094	209	25	g.	g.	PROPN
cana-6094	209	26	,	,	PUNCT
cana-6094	209	27	&	&	CCONJ
cana-6094	209	28	grohe	grohe	PROPN
cana-6094	209	29	,	,	PUNCT
cana-6094	209	30	m.	m.	NOUN
cana-6094	209	31	(	(	PUNCT
cana-6094	209	32	2019	2019	NUM
cana-6094	209	33	)	)	PUNCT
cana-6094	209	34	.	.	PUNCT
cana-6094	210	1	weisfeiler	weisfeiler	PROPN
cana-6094	210	2	and	and	CCONJ
cana-6094	210	3	leman	leman	PROPN
cana-6094	210	4	go	go	VERB
cana-6094	210	5	neural	neural	ADJ
cana-6094	210	6	:	:	PUNCT
cana-6094	210	7	higher	high	ADJ
cana-6094	210	8	-	-	PUNCT
cana-6094	210	9	order	order	NOUN
cana-6094	210	10	graph	graph	NOUN
cana-6094	210	11	neural	neural	ADJ
cana-6094	210	12	networks	network	NOUN
cana-6094	210	13	.	.	PUNCT
cana-6094	211	1	proceedings	proceeding	NOUN
cana-6094	211	2	of	of	ADP
cana-6094	211	3	the	the	DET
cana-6094	211	4	aaai	aaai	PROPN
cana-6094	211	5	conference	conference	NOUN
cana-6094	211	6	on	on	ADP
cana-6094	211	7	artificial	artificial	ADJ
cana-6094	211	8	intelligence	intelligence	NOUN
cana-6094	211	9	.	.	PUNCT
cana-6094	212	1	16	16	NUM
cana-6094	212	2	.	.	PUNCT
cana-6094	212	3	azizian	azizian	PROPN
cana-6094	212	4	,	,	PUNCT
cana-6094	212	5	w.	w.	PROPN
cana-6094	212	6	,	,	PUNCT
cana-6094	212	7	&	&	CCONJ
cana-6094	212	8	lelarge	lelarge	NOUN
cana-6094	212	9	,	,	PUNCT
cana-6094	212	10	m.	m.	NOUN
cana-6094	212	11	(	(	PUNCT
cana-6094	212	12	2021	2021	NUM
cana-6094	212	13	)	)	PUNCT
cana-6094	212	14	.	.	PUNCT
cana-6094	213	1	expressive	expressive	ADJ
cana-6094	213	2	power	power	NOUN
cana-6094	213	3	of	of	ADP
cana-6094	213	4	invariant	invariant	ADJ
cana-6094	213	5	and	and	CCONJ
cana-6094	213	6	equivariant	equivariant	ADJ
cana-6094	213	7	graph	graph	NOUN
cana-6094	213	8	neural	neural	ADJ
cana-6094	213	9	networks	network	NOUN
cana-6094	213	10	.	.	PUNCT
cana-6094	214	1	international	international	ADJ
cana-6094	214	2	conference	conference	NOUN
cana-6094	214	3	on	on	ADP
cana-6094	214	4	learning	learn	VERB
cana-6094	214	5	representations	representation	NOUN
cana-6094	214	6	(	(	PUNCT
cana-6094	214	7	iclr	iclr	NOUN
cana-6094	214	8	)	)	PUNCT
cana-6094	214	9	.	.	PUNCT
cana-6094	215	1	17	17	NUM
cana-6094	215	2	.	.	X
cana-6094	216	1	deac	deac	PROPN
cana-6094	216	2	,	,	PUNCT
cana-6094	216	3	a.	a.	NOUN
cana-6094	216	4	,	,	PUNCT
cana-6094	216	5	veličković	veličković	ADJ
cana-6094	216	6	,	,	PUNCT
cana-6094	216	7	p.	p.	NOUN
cana-6094	216	8	,	,	PUNCT
cana-6094	216	9	&	&	CCONJ
cana-6094	216	10	milinković	milinković	NOUN
cana-6094	216	11	,	,	PUNCT
cana-6094	216	12	o.	o.	NOUN
cana-6094	216	13	(	(	PUNCT
cana-6094	216	14	2019	2019	NUM
cana-6094	216	15	)	)	PUNCT
cana-6094	216	16	.	.	PUNCT
cana-6094	217	1	neural	neural	ADJ
cana-6094	217	2	reasoning	reasoning	NOUN
cana-6094	217	3	about	about	ADP
cana-6094	217	4	group	group	NOUN
cana-6094	217	5	structure	structure	NOUN
cana-6094	217	6	.	.	PUNCT
cana-6094	218	1	neurips	neurip	NOUN
cana-6094	218	2	workshop	workshop	VERB
cana-6094	218	3	on	on	ADP
cana-6094	218	4	graph	graph	NOUN
cana-6094	218	5	representation	representation	NOUN
cana-6094	218	6	learning	learning	NOUN
cana-6094	218	7	.	.	PUNCT
cana-6094	219	1	18	18	NUM
cana-6094	219	2	.	.	X
cana-6094	219	3	carr	carr	PROPN
cana-6094	219	4	,	,	PUNCT
cana-6094	219	5	s.	s.	PROPN
cana-6094	219	6	,	,	PUNCT
cana-6094	219	7	&	&	CCONJ
cana-6094	219	8	wu	wu	PROPN
cana-6094	219	9	,	,	PUNCT
cana-6094	219	10	j.	j.	PROPN
cana-6094	219	11	(	(	PUNCT
cana-6094	219	12	2022	2022	NUM
cana-6094	219	13	)	)	PUNCT
cana-6094	219	14	.	.	PUNCT
cana-6094	220	1	learning	learn	VERB
cana-6094	220	2	the	the	DET
cana-6094	220	3	irreducible	irreducible	ADJ
cana-6094	220	4	representations	representation	NOUN
cana-6094	220	5	of	of	ADP
cana-6094	220	6	finite	finite	ADJ
cana-6094	220	7	groups	group	NOUN
cana-6094	220	8	.	.	PUNCT
cana-6094	221	1	international	international	ADJ
cana-6094	221	2	conference	conference	NOUN
cana-6094	221	3	on	on	ADP
cana-6094	221	4	machine	machine	NOUN
cana-6094	221	5	learning	learning	NOUN
cana-6094	221	6	(	(	PUNCT
cana-6094	221	7	icml	icml	PROPN
cana-6094	221	8	)	)	PUNCT
cana-6094	221	9	.	.	PUNCT
cana-6094	222	1	19	19	NUM
cana-6094	222	2	.	.	X
cana-6094	222	3	ganea	ganea	NOUN
cana-6094	222	4	,	,	PUNCT
cana-6094	222	5	o.	o.	PROPN
cana-6094	222	6	e.	e.	PROPN
cana-6094	222	7	,	,	PUNCT
cana-6094	222	8	pattanaik	pattanaik	PROPN
cana-6094	222	9	,	,	PUNCT
cana-6094	222	10	l.	l.	PROPN
cana-6094	222	11	,	,	PUNCT
cana-6094	222	12	coley	coley	NOUN
cana-6094	222	13	,	,	PUNCT
cana-6094	222	14	c.	c.	PROPN
cana-6094	222	15	w.	w.	PROPN
cana-6094	222	16	,	,	PUNCT
cana-6094	222	17	barzilay	barzilay	PROPN
cana-6094	222	18	,	,	PUNCT
cana-6094	222	19	r.	r.	PROPN
cana-6094	222	20	,	,	PUNCT
cana-6094	222	21	jensen	jensen	PROPN
cana-6094	222	22	,	,	PUNCT
cana-6094	222	23	k.	k.	PROPN
cana-6094	222	24	f.	f.	PROPN
cana-6094	222	25	,	,	PUNCT
cana-6094	222	26	green	green	PROPN
cana-6094	222	27	,	,	PUNCT
cana-6094	222	28	w.	w.	PROPN
cana-6094	222	29	h.	h.	PROPN
cana-6094	222	30	,	,	PUNCT
cana-6094	222	31	&	&	CCONJ
cana-6094	222	32	jaakkola	jaakkola	PROPN
cana-6094	222	33	,	,	PUNCT
cana-6094	222	34	t.	t.	PROPN
cana-6094	222	35	s.	s.	PROPN
cana-6094	222	36	(	(	PUNCT
cana-6094	222	37	2021	2021	NUM
cana-6094	222	38	)	)	PUNCT
cana-6094	222	39	.	.	PUNCT
cana-6094	223	1	geomolecular	geomolecular	NOUN
cana-6094	223	2	:	:	PUNCT
cana-6094	223	3	a	a	DET
cana-6094	223	4	geometric	geometric	ADJ
cana-6094	223	5	deep	deep	ADJ
cana-6094	223	6	learning	learning	NOUN
cana-6094	223	7	library	library	NOUN
cana-6094	223	8	for	for	ADP
cana-6094	223	9	drug	drug	NOUN
cana-6094	223	10	discovery	discovery	NOUN
cana-6094	223	11	and	and	CCONJ
cana-6094	223	12	materials	material	NOUN
cana-6094	223	13	science	science	NOUN
cana-6094	223	14	.	.	PUNCT
cana-6094	224	1	journal	journal	PROPN
cana-6094	224	2	of	of	ADP
cana-6094	224	3	chemical	chemical	ADJ
cana-6094	224	4	information	information	NOUN
cana-6094	224	5	and	and	CCONJ
cana-6094	224	6	modeling	modeling	NOUN
cana-6094	224	7	.	.	PUNCT
cana-6094	225	1	20	20	NUM
cana-6094	225	2	.	.	PUNCT
cana-6094	226	1	liu	liu	PROPN
cana-6094	226	2	,	,	PUNCT
cana-6094	226	3	s.	s.	PROPN
cana-6094	226	4	,	,	PUNCT
cana-6094	226	5	&	&	CCONJ
cana-6094	226	6	wu	wu	PROPN
cana-6094	226	7	,	,	PUNCT
cana-6094	226	8	y.	y.	PROPN
cana-6094	226	9	(	(	PUNCT
cana-6094	226	10	2021	2021	NUM
cana-6094	226	11	)	)	PUNCT
cana-6094	226	12	.	.	PUNCT
cana-6094	227	1	learning	learn	VERB
cana-6094	227	2	algebraic	algebraic	ADJ
cana-6094	227	3	structures	structure	NOUN
cana-6094	227	4	:	:	PUNCT
cana-6094	227	5	towards	towards	ADP
cana-6094	227	6	a	a	DET
cana-6094	227	7	mathematical	mathematical	ADJ
cana-6094	227	8	ai	ai	NOUN
cana-6094	227	9	.	.	PUNCT
cana-6094	227	10	conference	conference	NOUN
cana-6094	227	11	on	on	ADP
cana-6094	227	12	automated	automate	VERB
cana-6094	227	13	deduction	deduction	NOUN
cana-6094	227	14	(	(	PUNCT
cana-6094	227	15	cade	cade	NOUN
cana-6094	227	16	)	)	PUNCT
cana-6094	227	17	.	.	PUNCT
cana-6094	228	1	21	21	NUM
cana-6094	228	2	.	.	PUNCT
cana-6094	229	1	joshi	joshi	PROPN
cana-6094	229	2	,	,	PUNCT
cana-6094	229	3	c.	c.	PROPN
cana-6094	229	4	k.	k.	PROPN
cana-6094	229	5	,	,	PUNCT
cana-6094	229	6	laurent	laurent	PROPN
cana-6094	229	7	,	,	PUNCT
cana-6094	229	8	t.	t.	PROPN
cana-6094	229	9	,	,	PUNCT
cana-6094	229	10	&	&	CCONJ
cana-6094	229	11	bresson	bresson	PROPN
cana-6094	229	12	,	,	PUNCT
cana-6094	229	13	x.	x.	NOUN
cana-6094	229	14	(	(	PUNCT
cana-6094	229	15	2019	2019	NUM
cana-6094	229	16	)	)	PUNCT
cana-6094	229	17	.	.	PUNCT
cana-6094	230	1	an	an	DET
cana-6094	230	2	efficient	efficient	ADJ
cana-6094	230	3	graph	graph	NOUN
cana-6094	230	4	convolutional	convolutional	ADJ
cana-6094	230	5	network	network	NOUN
cana-6094	230	6	technique	technique	NOUN
cana-6094	230	7	for	for	ADP
cana-6094	230	8	the	the	DET
cana-6094	230	9	travelling	travel	VERB
cana-6094	230	10	salesman	salesman	ADJ
cana-6094	230	11	problem	problem	NOUN
cana-6094	230	12	.	.	PUNCT
cana-6094	231	1	informs	inform	VERB
cana-6094	231	2	journal	journal	NOUN
cana-6094	231	3	on	on	ADP
cana-6094	231	4	computing	compute	VERB
cana-6094	231	5	.	.	PUNCT
cana-6094	232	1	22	22	NUM
cana-6094	232	2	.	.	PUNCT
cana-6094	233	1	khalil	khalil	PROPN
cana-6094	233	2	,	,	PUNCT
cana-6094	233	3	e.	e.	PROPN
cana-6094	233	4	,	,	PUNCT
cana-6094	233	5	dai	dai	PROPN
cana-6094	233	6	,	,	PUNCT
cana-6094	233	7	h.	h.	PROPN
cana-6094	233	8	,	,	PUNCT
cana-6094	233	9	zhang	zhang	PROPN
cana-6094	233	10	,	,	PUNCT
cana-6094	233	11	y.	y.	PROPN
cana-6094	233	12	,	,	PUNCT
cana-6094	233	13	dilkina	dilkina	PROPN
cana-6094	233	14	,	,	PUNCT
cana-6094	233	15	b.	b.	PROPN
cana-6094	233	16	,	,	PUNCT
cana-6094	233	17	&	&	CCONJ
cana-6094	233	18	song	song	PROPN
cana-6094	233	19	,	,	PUNCT
cana-6094	233	20	l.	l.	PROPN
cana-6094	233	21	(	(	PUNCT
cana-6094	233	22	2017	2017	NUM
cana-6094	233	23	)	)	PUNCT
cana-6094	233	24	.	.	PUNCT
cana-6094	234	1	learning	learn	VERB
cana-6094	234	2	combinatorial	combinatorial	ADJ
cana-6094	234	3	optimization	optimization	NOUN
cana-6094	234	4	algorithms	algorithm	NOUN
cana-6094	234	5	over	over	ADP
cana-6094	234	6	graphs	graph	NOUN
cana-6094	234	7	.	.	PUNCT
cana-6094	235	1	advances	advance	NOUN
cana-6094	235	2	in	in	ADP
cana-6094	235	3	neural	neural	ADJ
cana-6094	235	4	information	information	NOUN
cana-6094	235	5	processing	processing	NOUN
cana-6094	235	6	systems	system	NOUN
cana-6094	235	7	(	(	PUNCT
cana-6094	235	8	neurips	neurip	NOUN
cana-6094	235	9	)	)	PUNCT
cana-6094	235	10	.	.	PUNCT
cana-6094	236	1	23	23	NUM
cana-6094	236	2	.	.	PUNCT
cana-6094	236	3	prates	prate	NOUN
cana-6094	236	4	,	,	PUNCT
cana-6094	236	5	m.	m.	NOUN
cana-6094	236	6	,	,	PUNCT
cana-6094	236	7	avelar	avelar	PROPN
cana-6094	236	8	,	,	PUNCT
cana-6094	236	9	p.	p.	PROPN
cana-6094	236	10	h.	h.	PROPN
cana-6094	236	11	,	,	PUNCT
cana-6094	236	12	lemos	lemos	PROPN
cana-6094	236	13	,	,	PUNCT
cana-6094	236	14	h.	h.	PROPN
cana-6094	236	15	,	,	PUNCT
cana-6094	236	16	lamb	lamb	PROPN
cana-6094	236	17	,	,	PUNCT
cana-6094	236	18	l.	l.	PROPN
cana-6094	236	19	c.	c.	PROPN
cana-6094	236	20	,	,	PUNCT
cana-6094	236	21	&	&	CCONJ
cana-6094	236	22	vardi	vardi	PROPN
cana-6094	236	23	,	,	PUNCT
cana-6094	236	24	m.	m.	NOUN
cana-6094	236	25	y.	y.	PROPN
cana-6094	236	26	(	(	PUNCT
cana-6094	236	27	2019	2019	NUM
cana-6094	236	28	)	)	PUNCT
cana-6094	236	29	.	.	PUNCT
cana-6094	237	1	learning	learn	VERB
cana-6094	237	2	to	to	PART
cana-6094	237	3	solve	solve	VERB
cana-6094	237	4	np	np	ADV
cana-6094	237	5	-	-	PUNCT
cana-6094	237	6	complete	complete	ADJ
cana-6094	237	7	problems	problem	NOUN
cana-6094	237	8	:	:	PUNCT
cana-6094	237	9	a	a	DET
cana-6094	237	10	graph	graph	NOUN
cana-6094	237	11	neural	neural	ADJ
cana-6094	237	12	network	network	NOUN
cana-6094	237	13	for	for	ADP
cana-6094	237	14	decision	decision	NOUN
cana-6094	237	15	tsp	tsp	NOUN
cana-6094	237	16	.	.	PUNCT
cana-6094	238	1	proceedings	proceeding	NOUN
cana-6094	238	2	of	of	ADP
cana-6094	238	3	the	the	DET
cana-6094	238	4	aaai	aaai	PROPN
cana-6094	238	5	conference	conference	NOUN
cana-6094	238	6	on	on	ADP
cana-6094	238	7	artificial	artificial	ADJ
cana-6094	238	8	intelligence	intelligence	NOUN
cana-6094	238	9	.	.	PUNCT
cana-6094	239	1	24	24	NUM
cana-6094	239	2	.	.	PUNCT
cana-6094	239	3	bodnar	bodnar	PROPN
cana-6094	239	4	,	,	PUNCT
cana-6094	239	5	c.	c.	NOUN
cana-6094	239	6	,	,	PUNCT
cana-6094	239	7	frasca	frasca	PROPN
cana-6094	239	8	,	,	PUNCT
cana-6094	239	9	f.	f.	PROPN
cana-6094	239	10	,	,	PUNCT
cana-6094	239	11	wang	wang	PROPN
cana-6094	239	12	,	,	PUNCT
cana-6094	239	13	y.	y.	PROPN
cana-6094	239	14	,	,	PUNCT
cana-6094	239	15	otter	otter	NOUN
cana-6094	239	16	,	,	PUNCT
cana-6094	239	17	n.	n.	NOUN
cana-6094	239	18	,	,	PUNCT
cana-6094	239	19	montufar	montufar	PROPN
cana-6094	239	20	,	,	PUNCT
cana-6094	239	21	g.	g.	PROPN
cana-6094	239	22	f.	f.	PROPN
cana-6094	239	23	,	,	PUNCT
cana-6094	239	24	lio	lio	PROPN
cana-6094	239	25	,	,	PUNCT
cana-6094	239	26	p.	p.	PROPN
cana-6094	239	27	,	,	PUNCT
cana-6094	239	28	&	&	CCONJ
cana-6094	239	29	bronstein	bronstein	PROPN
cana-6094	239	30	,	,	PUNCT
cana-6094	239	31	m.	m.	NOUN
cana-6094	239	32	m.	m.	NOUN
cana-6094	239	33	(	(	PUNCT
cana-6094	239	34	2021	2021	NUM
cana-6094	239	35	)	)	PUNCT
cana-6094	239	36	.	.	PUNCT
cana-6094	240	1	weisfeiler	weisfeiler	PROPN
cana-6094	240	2	and	and	CCONJ
cana-6094	240	3	lehman	lehman	PROPN
cana-6094	240	4	go	go	VERB
cana-6094	240	5	topological	topological	ADJ
cana-6094	240	6	:	:	PUNCT
cana-6094	240	7	message	message	NOUN
cana-6094	240	8	passing	pass	VERB
cana-6094	240	9	simplicial	simplicial	ADJ
cana-6094	240	10	networks	network	NOUN
cana-6094	240	11	.	.	PUNCT
cana-6094	241	1	international	international	ADJ
cana-6094	241	2	conference	conference	NOUN
cana-6094	241	3	on	on	ADP
cana-6094	241	4	machine	machine	NOUN
cana-6094	241	5	learning	learning	NOUN
cana-6094	241	6	(	(	PUNCT
cana-6094	241	7	icml	icml	PROPN
cana-6094	241	8	)	)	PUNCT
cana-6094	241	9	.	.	PUNCT
cana-6094	242	1	25	25	NUM
cana-6094	242	2	.	.	X
cana-6094	243	1	horn	horn	NOUN
cana-6094	243	2	,	,	PUNCT
cana-6094	243	3	m.	m.	NOUN
cana-6094	243	4	,	,	PUNCT
cana-6094	243	5	de	de	PROPN
cana-6094	243	6	,	,	PUNCT
cana-6094	243	7	s.	s.	PROPN
cana-6094	243	8	,	,	PUNCT
cana-6094	243	9	&	&	CCONJ
cana-6094	243	10	berger	berger	PROPN
cana-6094	243	11	,	,	PUNCT
cana-6094	243	12	m.	m.	NOUN
cana-6094	243	13	(	(	PUNCT
cana-6094	243	14	2022	2022	NUM
cana-6094	243	15	)	)	PUNCT
cana-6094	243	16	.	.	PUNCT
cana-6094	244	1	topological	topological	ADJ
cana-6094	244	2	graph	graph	NOUN
cana-6094	244	3	neural	neural	ADJ
cana-6094	244	4	networks	network	NOUN
cana-6094	244	5	for	for	ADP
cana-6094	244	6	3d	3d	NUM
cana-6094	244	7	object	object	NOUN
cana-6094	244	8	recognition	recognition	NOUN
cana-6094	244	9	.	.	PUNCT
cana-6094	245	1	international	international	ADJ
cana-6094	245	2	conference	conference	NOUN
cana-6094	245	3	on	on	ADP
cana-6094	245	4	3d	3d	PROPN
cana-6094	245	5	vision	vision	NOUN
cana-6094	245	6	(	(	PUNCT
cana-6094	245	7	3dv	3dv	NOUN
cana-6094	245	8	)	)	PUNCT
cana-6094	245	9	.	.	PUNCT
cana-6094	246	1	26	26	NUM
cana-6094	246	2	.	.	PUNCT
cana-6094	247	1	poulenard	poulenard	NOUN
cana-6094	247	2	,	,	PUNCT
cana-6094	247	3	a.	a.	PROPN
cana-6094	247	4	,	,	PUNCT
cana-6094	247	5	&	&	CCONJ
cana-6094	247	6	ovsjanikov	ovsjanikov	PROPN
cana-6094	247	7	,	,	PUNCT
cana-6094	247	8	m.	m.	NOUN
cana-6094	247	9	(	(	PUNCT
cana-6094	247	10	2018	2018	NUM
cana-6094	247	11	)	)	PUNCT
cana-6094	247	12	.	.	PUNCT
cana-6094	248	1	multi	multi	ADJ
cana-6094	248	2	-	-	ADJ
cana-6094	248	3	directional	directional	ADJ
cana-6094	248	4	geodesic	geodesic	ADJ
cana-6094	248	5	neural	neural	ADJ
cana-6094	248	6	networks	network	NOUN
cana-6094	248	7	via	via	ADP
cana-6094	248	8	equivariant	equivariant	ADJ
cana-6094	248	9	convolution	convolution	NOUN
cana-6094	248	10	.	.	PUNCT
cana-6094	249	1	acm	acm	PROPN
cana-6094	249	2	transactions	transaction	NOUN
cana-6094	249	3	on	on	ADP
cana-6094	249	4	graphics	graphic	NOUN
cana-6094	249	5	(	(	PUNCT
cana-6094	249	6	tog	tog	NOUN
cana-6094	249	7	)	)	PUNCT
cana-6094	249	8	.	.	PUNCT
cana-6094	250	1	27	27	NUM
cana-6094	250	2	.	.	PUNCT
cana-6094	251	1	hofer	hofer	PROPN
cana-6094	251	2	,	,	PUNCT
cana-6094	251	3	c.	c.	PROPN
cana-6094	251	4	,	,	PUNCT
cana-6094	251	5	graf	graf	PROPN
cana-6094	251	6	,	,	PUNCT
cana-6094	251	7	f.	f.	PROPN
cana-6094	251	8	,	,	PUNCT
cana-6094	251	9	rieck	rieck	PROPN
cana-6094	251	10	,	,	PUNCT
cana-6094	251	11	b.	b.	PROPN
cana-6094	251	12	,	,	PUNCT
cana-6094	251	13	niethammer	niethammer	NOUN
cana-6094	251	14	,	,	PUNCT
cana-6094	251	15	m.	m.	NOUN
cana-6094	251	16	,	,	PUNCT
cana-6094	251	17	&	&	CCONJ
cana-6094	251	18	kwitt	kwitt	PROPN
cana-6094	251	19	,	,	PUNCT
cana-6094	251	20	r.	r.	PROPN
cana-6094	251	21	(	(	PUNCT
cana-6094	251	22	2020	2020	NUM
cana-6094	251	23	)	)	PUNCT
cana-6094	251	24	.	.	PUNCT
cana-6094	252	1	graph	graph	NOUN
cana-6094	252	2	filtration	filtration	NOUN
cana-6094	252	3	learning	learning	NOUN
cana-6094	252	4	.	.	PUNCT
cana-6094	253	1	international	international	ADJ
cana-6094	253	2	conference	conference	NOUN
cana-6094	253	3	on	on	ADP
cana-6094	253	4	machine	machine	NOUN
cana-6094	253	5	learning	learning	NOUN
cana-6094	253	6	(	(	PUNCT
cana-6094	253	7	icml	icml	PROPN
cana-6094	253	8	)	)	PUNCT
cana-6094	253	9	.	.	PUNCT
cana-6094	254	1	communications	communication	NOUN
cana-6094	254	2	on	on	ADP
cana-6094	254	3	applied	apply	VERB
cana-6094	254	4	nonlinear	nonlinear	ADJ
cana-6094	254	5	analysis	analysis	NOUN
cana-6094	254	6	issn	issn	NOUN
cana-6094	254	7	:	:	PUNCT
cana-6094	254	8	1074	1074	NUM
cana-6094	254	9	-	-	PUNCT
cana-6094	254	10	133x	133x	NUM
cana-6094	254	11	vol	vol	NOUN
cana-6094	254	12	29	29	NUM
cana-6094	254	13	no	no	NOUN
cana-6094	254	14	.	.	NOUN
cana-6094	254	15	4	4	NUM
cana-6094	254	16	(	(	PUNCT
cana-6094	254	17	2022	2022	NUM
cana-6094	254	18	)	)	PUNCT
cana-6094	254	19	38	38	NUM
cana-6094	254	20	https://internationalpubls.com	https://internationalpubls.com	SYM
cana-6094	254	21	28	28	NUM
cana-6094	254	22	.	.	PUNCT
cana-6094	255	1	bunch	bunch	PROPN
cana-6094	255	2	,	,	PUNCT
cana-6094	255	3	e.	e.	PROPN
cana-6094	255	4	,	,	PUNCT
cana-6094	255	5	you	you	PRON
cana-6094	255	6	,	,	PUNCT
cana-6094	255	7	q.	q.	PROPN
cana-6094	255	8	,	,	PUNCT
cana-6094	255	9	&	&	CCONJ
cana-6094	255	10	fung	fung	PROPN
cana-6094	255	11	,	,	PUNCT
cana-6094	255	12	g.	g.	PROPN
cana-6094	255	13	(	(	PUNCT
cana-6094	255	14	2020	2020	NUM
cana-6094	255	15	)	)	PUNCT
cana-6094	255	16	.	.	PUNCT
cana-6094	256	1	simplicial	simplicial	ADJ
cana-6094	256	2	2	2	NUM
cana-6094	256	3	-	-	PUNCT
cana-6094	256	4	complex	complex	ADJ
cana-6094	256	5	convolutional	convolutional	ADJ
cana-6094	256	6	neural	neural	ADJ
cana-6094	256	7	networks	network	NOUN
cana-6094	256	8	.	.	PUNCT
cana-6094	257	1	neurips	neurip	NOUN
cana-6094	257	2	workshop	workshop	NOUN
cana-6094	257	3	on	on	ADP
cana-6094	257	4	topological	topological	ADJ
cana-6094	257	5	data	datum	NOUN
cana-6094	257	6	analysis	analysis	NOUN
cana-6094	257	7	and	and	CCONJ
cana-6094	257	8	beyond	beyond	ADP
cana-6094	257	9	.	.	PROPN
cana-6094	258	1	29	29	NUM
cana-6094	258	2	.	.	X
cana-6094	259	1	pfaff	pfaff	PROPN
cana-6094	259	2	,	,	PUNCT
cana-6094	259	3	t.	t.	PROPN
cana-6094	259	4	,	,	PUNCT
cana-6094	259	5	fortunato	fortunato	PROPN
cana-6094	259	6	,	,	PUNCT
cana-6094	259	7	m.	m.	NOUN
cana-6094	259	8	,	,	PUNCT
cana-6094	259	9	sanchez	sanchez	PROPN
cana-6094	259	10	-	-	PUNCT
cana-6094	259	11	gonzalez	gonzalez	PROPN
cana-6094	259	12	,	,	PUNCT
cana-6094	259	13	a.	a.	PROPN
cana-6094	259	14	,	,	PUNCT
cana-6094	259	15	&	&	CCONJ
cana-6094	259	16	battaglia	battaglia	PROPN
cana-6094	259	17	,	,	PUNCT
cana-6094	259	18	p.	p.	NOUN
cana-6094	259	19	w.	w.	PROPN
cana-6094	259	20	(	(	PUNCT
cana-6094	259	21	2021	2021	NUM
cana-6094	259	22	)	)	PUNCT
cana-6094	259	23	.	.	PUNCT
cana-6094	260	1	learning	learn	VERB
cana-6094	260	2	mesh	mesh	NOUN
cana-6094	260	3	-	-	PUNCT
cana-6094	260	4	based	base	VERB
cana-6094	260	5	simulation	simulation	NOUN
cana-6094	260	6	with	with	ADP
cana-6094	260	7	graph	graph	NOUN
cana-6094	260	8	networks	network	NOUN
cana-6094	260	9	.	.	PUNCT
cana-6094	261	1	international	international	ADJ
cana-6094	261	2	conference	conference	NOUN
cana-6094	261	3	on	on	ADP
cana-6094	261	4	learning	learn	VERB
cana-6094	261	5	representations	representation	NOUN
cana-6094	261	6	(	(	PUNCT
cana-6094	261	7	iclr	iclr	NOUN
cana-6094	261	8	)	)	PUNCT
cana-6094	261	9	.	.	PUNCT
cana-6094	262	1	30	30	NUM
cana-6094	262	2	.	.	X
cana-6094	262	3	batzner	batzner	NOUN
cana-6094	262	4	,	,	PUNCT
cana-6094	262	5	s.	s.	PROPN
cana-6094	262	6	,	,	PUNCT
cana-6094	262	7	musaelian	musaelian	NOUN
cana-6094	262	8	,	,	PUNCT
cana-6094	262	9	a.	a.	NOUN
cana-6094	262	10	,	,	PUNCT
cana-6094	262	11	sun	sun	PROPN
cana-6094	262	12	,	,	PUNCT
cana-6094	262	13	l.	l.	PROPN
cana-6094	262	14	,	,	PUNCT
cana-6094	262	15	geiger	geiger	PROPN
cana-6094	262	16	,	,	PUNCT
cana-6094	262	17	m.	m.	NOUN
cana-6094	262	18	,	,	PUNCT
cana-6094	262	19	mailoa	mailoa	NOUN
cana-6094	262	20	,	,	PUNCT
cana-6094	262	21	j.	j.	PROPN
cana-6094	262	22	p.	p.	PROPN
cana-6094	262	23	,	,	PUNCT
cana-6094	262	24	kornbluth	kornbluth	PROPN
cana-6094	262	25	,	,	PUNCT
cana-6094	262	26	m.	m.	NOUN
cana-6094	262	27	,	,	PUNCT
cana-6094	262	28	...	...	PUNCT
cana-6094	262	29	&	&	CCONJ
cana-6094	262	30	kozinsky	kozinsky	PROPN
cana-6094	262	31	,	,	PUNCT
cana-6094	262	32	b.	b.	PROPN
cana-6094	262	33	(	(	PUNCT
cana-6094	262	34	2022	2022	NUM
cana-6094	262	35	)	)	PUNCT
cana-6094	262	36	.	.	PUNCT
cana-6094	263	1	e(3)equivariant	e(3)equivariant	NOUN
cana-6094	263	2	graph	graph	NOUN
cana-6094	263	3	neural	neural	ADJ
cana-6094	263	4	networks	network	NOUN
cana-6094	263	5	for	for	ADP
cana-6094	263	6	data	data	NOUN
cana-6094	263	7	-	-	PUNCT
cana-6094	263	8	efficient	efficient	ADJ
cana-6094	263	9	and	and	CCONJ
cana-6094	263	10	accurate	accurate	ADJ
cana-6094	263	11	interatomic	interatomic	ADJ
cana-6094	263	12	potentials	potential	NOUN
cana-6094	263	13	.	.	PUNCT
cana-6094	264	1	nature	nature	NOUN
cana-6094	264	2	communications	communication	NOUN
cana-6094	264	3	.	.	PUNCT
cana-6094	265	1	31	31	NUM
cana-6094	265	2	.	.	PUNCT
cana-6094	266	1	thölke	thölke	PROPN
cana-6094	266	2	,	,	PUNCT
cana-6094	266	3	p.	p.	NOUN
cana-6094	266	4	,	,	PUNCT
cana-6094	266	5	&	&	CCONJ
cana-6094	266	6	de	de	X
cana-6094	266	7	fabritiis	fabritiis	NOUN
cana-6094	266	8	,	,	PUNCT
cana-6094	266	9	g.	g.	PROPN
cana-6094	266	10	(	(	PUNCT
cana-6094	266	11	2022	2022	NUM
cana-6094	266	12	)	)	PUNCT
cana-6094	266	13	.	.	PUNCT
cana-6094	267	1	equivariant	equivariant	ADJ
cana-6094	267	2	transformers	transformer	NOUN
cana-6094	267	3	for	for	ADP
cana-6094	267	4	neural	neural	ADJ
cana-6094	267	5	network	network	NOUN
cana-6094	267	6	based	base	VERB
cana-6094	267	7	molecular	molecular	ADJ
cana-6094	267	8	potentials	potential	NOUN
cana-6094	267	9	.	.	PUNCT
cana-6094	268	1	international	international	ADJ
cana-6094	268	2	conference	conference	NOUN
cana-6094	268	3	on	on	ADP
cana-6094	268	4	learning	learn	VERB
cana-6094	268	5	representations	representation	NOUN
cana-6094	268	6	(	(	PUNCT
cana-6094	268	7	iclr	iclr	NOUN
cana-6094	268	8	)	)	PUNCT
cana-6094	268	9	.	.	PUNCT
cana-6094	269	1	32	32	NUM
cana-6094	269	2	.	.	PUNCT
cana-6094	270	1	cranmer	cranmer	PROPN
cana-6094	270	2	,	,	PUNCT
cana-6094	270	3	m.	m.	NOUN
cana-6094	270	4	,	,	PUNCT
cana-6094	270	5	sanchez	sanchez	PROPN
cana-6094	270	6	-	-	PUNCT
cana-6094	270	7	gonzalez	gonzalez	PROPN
cana-6094	270	8	,	,	PUNCT
cana-6094	270	9	a.	a.	PROPN
cana-6094	270	10	,	,	PUNCT
cana-6094	270	11	battaglia	battaglia	PROPN
cana-6094	270	12	,	,	PUNCT
cana-6094	270	13	p.	p.	PROPN
cana-6094	270	14	,	,	PUNCT
cana-6094	270	15	xu	xu	PROPN
cana-6094	270	16	,	,	PUNCT
cana-6094	270	17	r.	r.	PROPN
cana-6094	270	18	,	,	PUNCT
cana-6094	270	19	cranmer	cranmer	PROPN
cana-6094	270	20	,	,	PUNCT
cana-6094	270	21	k.	k.	PROPN
cana-6094	270	22	,	,	PUNCT
cana-6094	270	23	spergel	spergel	NOUN
cana-6094	270	24	,	,	PUNCT
cana-6094	270	25	d.	d.	PROPN
cana-6094	270	26	,	,	PUNCT
cana-6094	270	27	&	&	CCONJ
cana-6094	270	28	ho	ho	PROPN
cana-6094	270	29	,	,	PUNCT
cana-6094	270	30	s.	s.	PROPN
cana-6094	270	31	(	(	PUNCT
cana-6094	270	32	2021	2021	NUM
cana-6094	270	33	)	)	PUNCT
cana-6094	270	34	.	.	PUNCT
cana-6094	271	1	discovering	discover	VERB
cana-6094	271	2	symbolic	symbolic	ADJ
cana-6094	271	3	models	model	NOUN
cana-6094	271	4	from	from	ADP
cana-6094	271	5	deep	deep	ADJ
cana-6094	271	6	learning	learning	NOUN
cana-6094	271	7	with	with	ADP
cana-6094	271	8	inductive	inductive	ADJ
cana-6094	271	9	biases	bias	NOUN
cana-6094	271	10	.	.	PUNCT
cana-6094	272	1	advances	advance	NOUN
cana-6094	272	2	in	in	ADP
cana-6094	272	3	neural	neural	ADJ
cana-6094	272	4	information	information	NOUN
cana-6094	272	5	processing	processing	NOUN
cana-6094	272	6	systems	system	NOUN
cana-6094	272	7	(	(	PUNCT
cana-6094	272	8	neurips	neurip	NOUN
cana-6094	272	9	)	)	PUNCT
cana-6094	272	10	.	.	PUNCT
cana-6094	273	1	33	33	NUM
cana-6094	273	2	.	.	PUNCT
cana-6094	273	3	paliwal	paliwal	NOUN
cana-6094	273	4	,	,	PUNCT
cana-6094	273	5	a.	a.	NOUN
cana-6094	273	6	,	,	PUNCT
cana-6094	273	7	loos	loo	NOUN
cana-6094	273	8	,	,	PUNCT
cana-6094	273	9	s.	s.	PROPN
cana-6094	273	10	,	,	PUNCT
cana-6094	273	11	rabe	rabe	NOUN
cana-6094	273	12	,	,	PUNCT
cana-6094	273	13	m.	m.	NOUN
cana-6094	273	14	,	,	PUNCT
cana-6094	273	15	bansal	bansal	NOUN
cana-6094	273	16	,	,	PUNCT
cana-6094	273	17	k.	k.	PROPN
cana-6094	273	18	,	,	PUNCT
cana-6094	273	19	&	&	CCONJ
cana-6094	273	20	szegedy	szegedy	PROPN
cana-6094	273	21	,	,	PUNCT
cana-6094	273	22	c.	c.	PROPN
cana-6094	273	23	(	(	PUNCT
cana-6094	273	24	2020	2020	NUM
cana-6094	273	25	)	)	PUNCT
cana-6094	273	26	.	.	PUNCT
cana-6094	274	1	graph	graph	NOUN
cana-6094	274	2	representations	representation	NOUN
cana-6094	274	3	for	for	ADP
cana-6094	274	4	higher	high	ADJ
cana-6094	274	5	-	-	PUNCT
cana-6094	274	6	order	order	NOUN
cana-6094	274	7	logic	logic	NOUN
cana-6094	274	8	and	and	CCONJ
cana-6094	274	9	theorem	theorem	ADJ
cana-6094	274	10	proving	proving	NOUN
cana-6094	274	11	.	.	PUNCT
cana-6094	275	1	proceedings	proceeding	NOUN
cana-6094	275	2	of	of	ADP
cana-6094	275	3	the	the	DET
cana-6094	275	4	aaai	aaai	PROPN
cana-6094	275	5	conference	conference	NOUN
cana-6094	275	6	on	on	ADP
cana-6094	275	7	artificial	artificial	ADJ
cana-6094	275	8	intelligence	intelligence	NOUN
cana-6094	275	9	.	.	PUNCT
cana-6094	276	1	34	34	NUM
cana-6094	276	2	.	.	PUNCT
cana-6094	276	3	loos	loos	PROPN
cana-6094	276	4	,	,	PUNCT
cana-6094	276	5	s.	s.	PROPN
cana-6094	276	6	,	,	PUNCT
cana-6094	276	7	irving	irving	PROPN
cana-6094	276	8	,	,	PUNCT
cana-6094	276	9	g.	g.	PROPN
cana-6094	276	10	,	,	PUNCT
cana-6094	276	11	szegedy	szegedy	PROPN
cana-6094	276	12	,	,	PUNCT
cana-6094	276	13	c.	c.	NOUN
cana-6094	276	14	,	,	PUNCT
cana-6094	276	15	&	&	CCONJ
cana-6094	276	16	kaliszyk	kaliszyk	PROPN
cana-6094	276	17	,	,	PUNCT
cana-6094	276	18	c.	c.	NOUN
cana-6094	276	19	(	(	PUNCT
cana-6094	276	20	2017	2017	NUM
cana-6094	276	21	)	)	PUNCT
cana-6094	276	22	.	.	PUNCT
cana-6094	277	1	deep	deep	ADJ
cana-6094	277	2	network	network	NOUN
cana-6094	277	3	guided	guide	VERB
cana-6094	277	4	proof	proof	NOUN
cana-6094	277	5	search	search	NOUN
cana-6094	277	6	.	.	PUNCT
cana-6094	278	1	international	international	ADJ
cana-6094	278	2	joint	joint	ADJ
cana-6094	278	3	conference	conference	NOUN
cana-6094	278	4	on	on	ADP
cana-6094	278	5	automated	automate	VERB
cana-6094	278	6	reasoning	reasoning	NOUN
cana-6094	278	7	(	(	PUNCT
cana-6094	278	8	ijcar	ijcar	NOUN
cana-6094	278	9	)	)	PUNCT
cana-6094	278	10	.	.	PUNCT
cana-6094	279	1	35	35	NUM
cana-6094	279	2	.	.	PUNCT
cana-6094	279	3	wu	wu	PROPN
cana-6094	279	4	,	,	PUNCT
cana-6094	279	5	m.	m.	NOUN
cana-6094	279	6	,	,	PUNCT
cana-6094	279	7	norrish	norrish	ADJ
cana-6094	279	8	,	,	PUNCT
cana-6094	279	9	m.	m.	NOUN
cana-6094	279	10	,	,	PUNCT
cana-6094	279	11	walder	walder	NOUN
cana-6094	279	12	,	,	PUNCT
cana-6094	279	13	c.	c.	PROPN
cana-6094	279	14	,	,	PUNCT
cana-6094	279	15	&	&	CCONJ
cana-6094	279	16	dezfouli	dezfouli	PROPN
cana-6094	279	17	,	,	PUNCT
cana-6094	279	18	a.	a.	NOUN
cana-6094	279	19	(	(	PUNCT
cana-6094	279	20	2021	2021	NUM
cana-6094	279	21	)	)	PUNCT
cana-6094	279	22	.	.	PUNCT
cana-6094	280	1	tactic	tactic	ADJ
cana-6094	280	2	prediction	prediction	NOUN
cana-6094	280	3	and	and	CCONJ
cana-6094	280	4	generation	generation	NOUN
cana-6094	280	5	for	for	ADP
cana-6094	280	6	interactive	interactive	ADJ
cana-6094	280	7	theorem	theorem	ADJ
cana-6094	280	8	proving	proving	NOUN
cana-6094	280	9	.	.	PUNCT
cana-6094	281	1	conference	conference	NOUN
cana-6094	281	2	on	on	ADP
cana-6094	281	3	automated	automate	VERB
cana-6094	281	4	deduction	deduction	NOUN
cana-6094	281	5	(	(	PUNCT
cana-6094	281	6	cade	cade	NOUN
cana-6094	281	7	)	)	PUNCT
cana-6094	281	8	.	.	PUNCT
cana-6094	282	1	36	36	NUM
cana-6094	282	2	.	.	PUNCT
cana-6094	282	3	rabe	rabe	NOUN
cana-6094	282	4	,	,	PUNCT
cana-6094	282	5	m.	m.	NOUN
cana-6094	282	6	n.	n.	PROPN
cana-6094	282	7	,	,	PUNCT
cana-6094	282	8	&	&	CCONJ
cana-6094	282	9	szegedy	szegedy	PROPN
cana-6094	282	10	,	,	PUNCT
cana-6094	282	11	c.	c.	NOUN
cana-6094	282	12	(	(	PUNCT
cana-6094	282	13	2021	2021	NUM
cana-6094	282	14	)	)	PUNCT
cana-6094	282	15	.	.	PUNCT
cana-6094	283	1	the	the	DET
cana-6094	283	2	graph	graph	NOUN
cana-6094	283	3	of	of	ADP
cana-6094	283	4	theorems	theorem	NOUN
cana-6094	283	5	.	.	PUNCT
cana-6094	284	1	conference	conference	NOUN
cana-6094	284	2	on	on	ADP
cana-6094	284	3	automated	automate	VERB
cana-6094	284	4	deduction	deduction	NOUN
cana-6094	284	5	(	(	PUNCT
cana-6094	284	6	cade	cade	NOUN
cana-6094	284	7	)	)	PUNCT
cana-6094	284	8	.	.	PUNCT
cana-6094	285	1	37	37	NUM
cana-6094	285	2	.	.	NOUN
cana-6094	285	3	39	39	NUM
cana-6094	285	4	.	.	PUNCT
cana-6094	286	1	you	you	PRON
cana-6094	286	2	,	,	PUNCT
cana-6094	286	3	j.	j.	PROPN
cana-6094	286	4	,	,	PUNCT
cana-6094	286	5	ying	ying	PROPN
cana-6094	286	6	,	,	PUNCT
cana-6094	286	7	r.	r.	PROPN
cana-6094	286	8	,	,	PUNCT
cana-6094	286	9	ren	ren	PROPN
cana-6094	286	10	,	,	PUNCT
cana-6094	286	11	x.	x.	PROPN
cana-6094	286	12	,	,	PUNCT
cana-6094	286	13	hamilton	hamilton	PROPN
cana-6094	286	14	,	,	PUNCT
cana-6094	286	15	w.	w.	PROPN
cana-6094	286	16	l.	l.	PROPN
cana-6094	286	17	,	,	PUNCT
cana-6094	286	18	&	&	CCONJ
cana-6094	286	19	leskovec	leskovec	PROPN
cana-6094	286	20	,	,	PUNCT
cana-6094	286	21	j.	j.	PROPN
cana-6094	286	22	(	(	PUNCT
cana-6094	286	23	2018	2018	NUM
cana-6094	286	24	)	)	PUNCT
cana-6094	286	25	.	.	PUNCT
cana-6094	287	1	graphrnn	graphrnn	NOUN
cana-6094	287	2	:	:	PUNCT
cana-6094	287	3	generating	generate	VERB
cana-6094	287	4	realistic	realistic	ADJ
cana-6094	287	5	graphs	graph	NOUN
cana-6094	287	6	with	with	ADP
cana-6094	287	7	deep	deep	ADJ
cana-6094	287	8	auto	auto	NOUN
cana-6094	287	9	-	-	PUNCT
cana-6094	287	10	regressive	regressive	ADJ
cana-6094	287	11	models	model	NOUN
cana-6094	287	12	.	.	PUNCT
cana-6094	288	1	international	international	ADJ
cana-6094	288	2	conference	conference	NOUN
cana-6094	288	3	on	on	ADP
cana-6094	288	4	machine	machine	NOUN
cana-6094	288	5	learning	learning	NOUN
cana-6094	288	6	(	(	PUNCT
cana-6094	288	7	icml	icml	PROPN
cana-6094	288	8	)	)	PUNCT
cana-6094	288	9	.	.	PUNCT
cana-6094	289	1	38	38	NUM
cana-6094	289	2	.	.	PUNCT
cana-6094	289	3	liao	liao	PROPN
cana-6094	289	4	,	,	PUNCT
cana-6094	289	5	r.	r.	PROPN
cana-6094	289	6	,	,	PUNCT
cana-6094	289	7	li	li	PROPN
cana-6094	289	8	,	,	PUNCT
cana-6094	289	9	y.	y.	PROPN
cana-6094	289	10	,	,	PUNCT
cana-6094	289	11	song	song	NOUN
cana-6094	289	12	,	,	PUNCT
cana-6094	289	13	y.	y.	PROPN
cana-6094	289	14	,	,	PUNCT
cana-6094	289	15	wang	wang	PROPN
cana-6094	289	16	,	,	PUNCT
cana-6094	289	17	s.	s.	PROPN
cana-6094	289	18	,	,	PUNCT
cana-6094	289	19	hamilton	hamilton	PROPN
cana-6094	289	20	,	,	PUNCT
cana-6094	289	21	w.	w.	PROPN
cana-6094	289	22	l.	l.	PROPN
cana-6094	289	23	,	,	PUNCT
cana-6094	289	24	duvenaud	duvenaud	PROPN
cana-6094	289	25	,	,	PUNCT
cana-6094	289	26	d.	d.	PROPN
cana-6094	289	27	k.	k.	PROPN
cana-6094	289	28	,	,	PUNCT
cana-6094	289	29	...	...	PUNCT
cana-6094	289	30	&	&	CCONJ
cana-6094	289	31	urtasun	urtasun	PROPN
cana-6094	289	32	,	,	PUNCT
cana-6094	289	33	r.	r.	PROPN
cana-6094	289	34	(	(	PUNCT
cana-6094	289	35	2019	2019	NUM
cana-6094	289	36	)	)	PUNCT
cana-6094	289	37	.	.	PUNCT
cana-6094	290	1	efficient	efficient	ADJ
cana-6094	290	2	graph	graph	NOUN
cana-6094	290	3	generation	generation	NOUN
cana-6094	290	4	with	with	ADP
cana-6094	290	5	graph	graph	NOUN
cana-6094	290	6	recurrent	recurrent	ADJ
cana-6094	290	7	attention	attention	NOUN
cana-6094	290	8	networks	network	NOUN
cana-6094	290	9	.	.	PUNCT
cana-6094	291	1	advances	advance	NOUN
cana-6094	291	2	in	in	ADP
cana-6094	291	3	neural	neural	ADJ
cana-6094	291	4	information	information	NOUN
cana-6094	291	5	processing	processing	NOUN
cana-6094	291	6	systems	system	NOUN
cana-6094	291	7	(	(	PUNCT
cana-6094	291	8	neurips	neurip	NOUN
cana-6094	291	9	)	)	PUNCT
cana-6094	291	10	.	.	PUNCT
cana-6094	292	1	39	39	NUM
cana-6094	292	2	.	.	PUNCT
cana-6094	293	1	bojchevski	bojchevski	PROPN
cana-6094	293	2	,	,	PUNCT
cana-6094	293	3	a.	a.	NOUN
cana-6094	293	4	,	,	PUNCT
cana-6094	293	5	shchur	shchur	ADJ
cana-6094	293	6	,	,	PUNCT
cana-6094	293	7	o.	o.	PROPN
cana-6094	293	8	,	,	PUNCT
cana-6094	293	9	zügner	zügner	PROPN
cana-6094	293	10	,	,	PUNCT
cana-6094	293	11	d.	d.	PROPN
cana-6094	293	12	,	,	PUNCT
cana-6094	293	13	&	&	CCONJ
cana-6094	293	14	günnemann	günnemann	PROPN
cana-6094	293	15	,	,	PUNCT
cana-6094	293	16	s.	s.	PROPN
cana-6094	293	17	(	(	PUNCT
cana-6094	293	18	2018	2018	NUM
cana-6094	293	19	)	)	PUNCT
cana-6094	293	20	.	.	PUNCT
cana-6094	294	1	netgan	netgan	ADV
cana-6094	294	2	:	:	PUNCT
cana-6094	295	1	generating	generate	VERB
cana-6094	295	2	graphs	graph	NOUN
cana-6094	295	3	via	via	ADP
cana-6094	295	4	random	random	ADJ
cana-6094	295	5	walks	walk	NOUN
cana-6094	295	6	.	.	PUNCT
cana-6094	296	1	international	international	ADJ
cana-6094	296	2	conference	conference	NOUN
cana-6094	296	3	on	on	ADP
cana-6094	296	4	machine	machine	NOUN
cana-6094	296	5	learning	learning	NOUN
cana-6094	296	6	(	(	PUNCT
cana-6094	296	7	icml	icml	PROPN
cana-6094	296	8	)	)	PUNCT
cana-6094	296	9	.	.	PUNCT
cana-6094	297	1	40	40	NUM
cana-6094	297	2	.	.	PUNCT
cana-6094	298	1	ying	ying	PROPN
cana-6094	298	2	,	,	PUNCT
cana-6094	298	3	z.	z.	PROPN
cana-6094	298	4	,	,	PUNCT
cana-6094	298	5	bourgeois	bourgeois	PROPN
cana-6094	298	6	,	,	PUNCT
cana-6094	298	7	d.	d.	PROPN
cana-6094	298	8	,	,	PUNCT
cana-6094	298	9	you	you	PRON
cana-6094	298	10	,	,	PUNCT
cana-6094	298	11	j.	j.	PROPN
cana-6094	298	12	,	,	PUNCT
cana-6094	298	13	zitnik	zitnik	NOUN
cana-6094	298	14	,	,	PUNCT
cana-6094	298	15	m.	m.	NOUN
cana-6094	298	16	,	,	PUNCT
cana-6094	298	17	&	&	CCONJ
cana-6094	298	18	leskovec	leskovec	PROPN
cana-6094	298	19	,	,	PUNCT
cana-6094	298	20	j.	j.	PROPN
cana-6094	298	21	(	(	PUNCT
cana-6094	298	22	2019	2019	NUM
cana-6094	298	23	)	)	PUNCT
cana-6094	298	24	.	.	PUNCT
cana-6094	299	1	gnnexplainer	gnnexplainer	NOUN
cana-6094	299	2	:	:	PUNCT
cana-6094	299	3	generating	generate	VERB
cana-6094	299	4	explanations	explanation	NOUN
cana-6094	299	5	for	for	ADP
cana-6094	299	6	graph	graph	NOUN
cana-6094	299	7	neural	neural	ADJ
cana-6094	299	8	networks	network	NOUN
cana-6094	299	9	.	.	PUNCT
cana-6094	300	1	advances	advance	NOUN
cana-6094	300	2	in	in	ADP
cana-6094	300	3	neural	neural	ADJ
cana-6094	300	4	information	information	NOUN
cana-6094	300	5	processing	processing	NOUN
cana-6094	300	6	systems	system	NOUN
cana-6094	300	7	(	(	PUNCT
cana-6094	300	8	neurips	neurip	NOUN
cana-6094	300	9	)	)	PUNCT
cana-6094	300	10	.	.	PUNCT
cana-6094	301	1	41	41	NUM
cana-6094	301	2	.	.	PUNCT
cana-6094	302	1	garg	garg	NOUN
cana-6094	302	2	,	,	PUNCT
cana-6094	302	3	v.	v.	PROPN
cana-6094	302	4	,	,	PUNCT
cana-6094	302	5	jegelka	jegelka	PROPN
cana-6094	302	6	,	,	PUNCT
cana-6094	302	7	s.	s.	PROPN
cana-6094	302	8	,	,	PUNCT
cana-6094	302	9	&	&	CCONJ
cana-6094	302	10	jaakkola	jaakkola	PROPN
cana-6094	302	11	,	,	PUNCT
cana-6094	302	12	t.	t.	PROPN
cana-6094	302	13	(	(	PUNCT
cana-6094	302	14	2020	2020	NUM
cana-6094	302	15	)	)	PUNCT
cana-6094	302	16	.	.	PUNCT
cana-6094	303	1	generalization	generalization	NOUN
cana-6094	303	2	and	and	CCONJ
cana-6094	303	3	representational	representational	ADJ
cana-6094	303	4	limits	limit	NOUN
cana-6094	303	5	of	of	ADP
cana-6094	303	6	graph	graph	NOUN
cana-6094	303	7	neural	neural	ADJ
cana-6094	303	8	networks	network	NOUN
cana-6094	303	9	.	.	PUNCT
cana-6094	304	1	international	international	ADJ
cana-6094	304	2	conference	conference	NOUN
cana-6094	304	3	on	on	ADP
cana-6094	304	4	machine	machine	NOUN
cana-6094	304	5	learning	learning	NOUN
cana-6094	304	6	(	(	PUNCT
cana-6094	304	7	icml	icml	PROPN
cana-6094	304	8	)	)	PUNCT
cana-6094	304	9	.	.	PUNCT
cana-6094	305	1	42	42	NUM
cana-6094	305	2	.	.	X
cana-6094	306	1	oono	oono	PROPN
cana-6094	306	2	,	,	PUNCT
cana-6094	306	3	k.	k.	PROPN
cana-6094	306	4	,	,	PUNCT
cana-6094	306	5	&	&	CCONJ
cana-6094	306	6	suzuki	suzuki	PROPN
cana-6094	306	7	,	,	PUNCT
cana-6094	306	8	t.	t.	PROPN
cana-6094	306	9	(	(	PUNCT
cana-6094	306	10	2020	2020	NUM
cana-6094	306	11	)	)	PUNCT
cana-6094	306	12	.	.	PUNCT
cana-6094	307	1	graph	graph	NOUN
cana-6094	307	2	neural	neural	ADJ
cana-6094	307	3	networks	network	NOUN
cana-6094	307	4	exponentially	exponentially	ADV
cana-6094	307	5	lose	lose	VERB
cana-6094	307	6	expressive	expressive	ADJ
cana-6094	307	7	power	power	NOUN
cana-6094	307	8	for	for	ADP
cana-6094	307	9	node	node	ADJ
cana-6094	307	10	classification	classification	NOUN
cana-6094	307	11	.	.	PUNCT
cana-6094	308	1	international	international	ADJ
cana-6094	308	2	conference	conference	NOUN
cana-6094	308	3	on	on	ADP
cana-6094	308	4	learning	learn	VERB
cana-6094	308	5	representations	representation	NOUN
cana-6094	308	6	(	(	PUNCT
cana-6094	308	7	iclr	iclr	NOUN
cana-6094	308	8	)	)	PUNCT
cana-6094	308	9	.	.	PUNCT
cana-6094	309	1	43	43	NUM
cana-6094	309	2	.	.	PUNCT
cana-6094	310	1	chen	chen	PROPN
cana-6094	310	2	,	,	PUNCT
cana-6094	310	3	z.	z.	PROPN
cana-6094	310	4	,	,	PUNCT
cana-6094	310	5	villar	villar	PROPN
cana-6094	310	6	,	,	PUNCT
cana-6094	310	7	s.	s.	PROPN
cana-6094	310	8	,	,	PUNCT
cana-6094	310	9	chen	chen	PROPN
cana-6094	310	10	,	,	PUNCT
cana-6094	310	11	l.	l.	PROPN
cana-6094	310	12	,	,	PUNCT
cana-6094	310	13	&	&	CCONJ
cana-6094	310	14	bruna	bruna	PROPN
cana-6094	310	15	,	,	PUNCT
cana-6094	310	16	j.	j.	PROPN
cana-6094	310	17	(	(	PUNCT
cana-6094	310	18	2021	2021	NUM
cana-6094	310	19	)	)	PUNCT
cana-6094	310	20	.	.	PUNCT
cana-6094	311	1	on	on	ADP
cana-6094	311	2	the	the	DET
cana-6094	311	3	equivalence	equivalence	NOUN
cana-6094	311	4	between	between	ADP
cana-6094	311	5	graph	graph	NOUN
cana-6094	311	6	isomorphism	isomorphism	NOUN
cana-6094	311	7	testing	testing	NOUN
cana-6094	311	8	and	and	CCONJ
cana-6094	311	9	function	function	NOUN
cana-6094	311	10	approximation	approximation	NOUN
cana-6094	311	11	with	with	ADP
cana-6094	311	12	gnns	gnn	NOUN
cana-6094	311	13	.	.	PUNCT
cana-6094	312	1	advances	advance	NOUN
cana-6094	312	2	in	in	ADP
cana-6094	312	3	neural	neural	ADJ
cana-6094	312	4	information	information	NOUN
cana-6094	312	5	processing	processing	NOUN
cana-6094	312	6	systems	system	NOUN
cana-6094	312	7	(	(	PUNCT
cana-6094	312	8	neurips	neurip	NOUN
cana-6094	312	9	)	)	PUNCT
cana-6094	312	10	.	.	PUNCT
cana-6094	313	1	44	44	NUM
cana-6094	313	2	.	.	PUNCT
cana-6094	314	1	davies	davy	NOUN
cana-6094	314	2	,	,	PUNCT
cana-6094	314	3	a.	a.	NOUN
cana-6094	314	4	,	,	PUNCT
cana-6094	314	5	veličković	veličković	ADJ
cana-6094	314	6	,	,	PUNCT
cana-6094	314	7	p.	p.	NOUN
cana-6094	314	8	,	,	PUNCT
cana-6094	314	9	buesing	buese	VERB
cana-6094	314	10	,	,	PUNCT
cana-6094	314	11	l.	l.	PROPN
cana-6094	314	12	,	,	PUNCT
cana-6094	314	13	blackwell	blackwell	PROPN
cana-6094	314	14	,	,	PUNCT
cana-6094	314	15	s.	s.	PROPN
cana-6094	314	16	,	,	PUNCT
cana-6094	314	17	zheng	zheng	PROPN
cana-6094	314	18	,	,	PUNCT
cana-6094	314	19	d.	d.	PROPN
cana-6094	314	20	,	,	PUNCT
cana-6094	314	21	tomasev	tomasev	AUX
cana-6094	314	22	,	,	PUNCT
cana-6094	314	23	n.	n.	NOUN
cana-6094	314	24	,	,	PUNCT
cana-6094	314	25	...	...	PUNCT
cana-6094	314	26	&	&	CCONJ
cana-6094	314	27	williamson	williamson	PROPN
cana-6094	314	28	,	,	PUNCT
cana-6094	314	29	m.	m.	NOUN
cana-6094	314	30	(	(	PUNCT
cana-6094	314	31	2021	2021	NUM
cana-6094	314	32	)	)	PUNCT
cana-6094	314	33	.	.	PUNCT
cana-6094	315	1	advancing	advance	VERB
cana-6094	315	2	mathematics	mathematic	NOUN
cana-6094	315	3	by	by	ADP
cana-6094	315	4	guiding	guide	VERB
cana-6094	315	5	human	human	ADJ
cana-6094	315	6	intuition	intuition	NOUN
cana-6094	315	7	with	with	ADP
cana-6094	315	8	ai	ai	PROPN
cana-6094	315	9	.	.	PROPN
cana-6094	315	10	nature	nature	NOUN
cana-6094	315	11	,	,	PUNCT
cana-6094	315	12	*	*	NOUN
cana-6094	315	13	600*(7887	600*(7887	NUM
cana-6094	315	14	)	)	PUNCT
cana-6094	315	15	,	,	PUNCT
cana-6094	315	16	70	70	NUM
cana-6094	315	17	-	-	SYM
cana-6094	315	18	74	74	NUM
cana-6094	315	19	.	.	PUNCT
cana-6094	316	1	45	45	NUM
cana-6094	316	2	.	.	NUM
cana-6094	316	3	iten	iten	PROPN
cana-6094	316	4	,	,	PUNCT
cana-6094	316	5	r.	r.	PROPN
cana-6094	316	6	,	,	PUNCT
cana-6094	316	7	metger	metger	NOUN
cana-6094	316	8	,	,	PUNCT
cana-6094	316	9	t.	t.	NOUN
cana-6094	316	10	,	,	PUNCT
cana-6094	316	11	wilming	wilming	NOUN
cana-6094	316	12	,	,	PUNCT
cana-6094	316	13	h.	h.	PROPN
cana-6094	316	14	,	,	PUNCT
cana-6094	316	15	del	del	PROPN
cana-6094	316	16	rio	rio	PROPN
cana-6094	316	17	,	,	PUNCT
cana-6094	316	18	l.	l.	PROPN
cana-6094	316	19	,	,	PUNCT
cana-6094	316	20	&	&	CCONJ
cana-6094	316	21	renner	renner	PROPN
cana-6094	316	22	,	,	PUNCT
cana-6094	316	23	r.	r.	PROPN
cana-6094	316	24	(	(	PUNCT
cana-6094	316	25	2020	2020	NUM
cana-6094	316	26	)	)	PUNCT
cana-6094	316	27	.	.	PUNCT
cana-6094	317	1	discovering	discover	VERB
cana-6094	317	2	physical	physical	ADJ
cana-6094	317	3	concepts	concept	NOUN
cana-6094	317	4	with	with	ADP
cana-6094	317	5	neural	neural	ADJ
cana-6094	317	6	networks	network	NOUN
cana-6094	317	7	.	.	PUNCT
cana-6094	318	1	physical	physical	ADJ
cana-6094	318	2	review	review	NOUN
cana-6094	318	3	letters	letter	NOUN
cana-6094	318	4	.	.	PUNCT
cana-6094	319	1	communications	communication	NOUN
cana-6094	319	2	on	on	ADP
cana-6094	319	3	applied	apply	VERB
cana-6094	319	4	nonlinear	nonlinear	ADJ
cana-6094	319	5	analysis	analysis	NOUN
cana-6094	319	6	issn	issn	NOUN
cana-6094	319	7	:	:	PUNCT
cana-6094	319	8	1074	1074	NUM
cana-6094	319	9	-	-	PUNCT
cana-6094	319	10	133x	133x	NUM
cana-6094	319	11	vol	vol	NOUN
cana-6094	319	12	29	29	NUM
cana-6094	319	13	no	no	NOUN
cana-6094	319	14	.	.	NOUN
cana-6094	319	15	4	4	NUM
cana-6094	319	16	(	(	PUNCT
cana-6094	319	17	2022	2022	NUM
cana-6094	319	18	)	)	PUNCT
cana-6094	319	19	39	39	NUM
cana-6094	320	1	https://internationalpubls.com	https://internationalpubls.com	X
cana-6094	320	2	46	46	NUM
cana-6094	320	3	.	.	PUNCT
cana-6094	320	4	raissi	raissi	ADJ
cana-6094	320	5	,	,	PUNCT
cana-6094	320	6	m.	m.	NOUN
cana-6094	320	7	,	,	PUNCT
cana-6094	320	8	perdikaris	perdikaris	NOUN
cana-6094	320	9	,	,	PUNCT
cana-6094	320	10	p.	p.	NOUN
cana-6094	320	11	,	,	PUNCT
cana-6094	320	12	&	&	CCONJ
cana-6094	320	13	karniadakis	karniadakis	PROPN
cana-6094	320	14	,	,	PUNCT
cana-6094	320	15	g.	g.	PROPN
cana-6094	320	16	e.	e.	PROPN
cana-6094	320	17	(	(	PUNCT
cana-6094	320	18	2019	2019	NUM
cana-6094	320	19	)	)	PUNCT
cana-6094	320	20	.	.	PUNCT
cana-6094	321	1	physics	physics	NOUN
cana-6094	321	2	-	-	PUNCT
cana-6094	321	3	informed	inform	VERB
cana-6094	321	4	neural	neural	ADJ
cana-6094	321	5	networks	network	NOUN
cana-6094	321	6	:	:	PUNCT
cana-6094	321	7	a	a	DET
cana-6094	321	8	deep	deep	ADJ
cana-6094	321	9	learning	learning	NOUN
cana-6094	321	10	framework	framework	NOUN
cana-6094	321	11	for	for	ADP
cana-6094	321	12	solving	solve	VERB
cana-6094	321	13	forward	forward	ADV
cana-6094	321	14	and	and	CCONJ
cana-6094	321	15	inverse	inverse	NOUN
cana-6094	321	16	problems	problem	NOUN
cana-6094	321	17	involving	involve	VERB
cana-6094	321	18	nonlinear	nonlinear	ADJ
cana-6094	321	19	partial	partial	ADJ
cana-6094	321	20	differential	differential	NOUN
cana-6094	321	21	equations	equation	NOUN
cana-6094	321	22	.	.	PUNCT
cana-6094	322	1	journal	journal	NOUN
cana-6094	322	2	of	of	ADP
cana-6094	322	3	computational	computational	ADJ
cana-6094	322	4	physics	physic	NOUN
cana-6094	322	5	.	.	PUNCT
cana-6094	323	1	47	47	NUM
cana-6094	323	2	.	.	PUNCT
cana-6094	323	3	jumper	jumper	PROPN
cana-6094	323	4	,	,	PUNCT
cana-6094	323	5	j.	j.	PROPN
cana-6094	323	6	,	,	PUNCT
cana-6094	323	7	evans	evans	PROPN
cana-6094	323	8	,	,	PUNCT
cana-6094	323	9	r.	r.	PROPN
cana-6094	323	10	,	,	PUNCT
cana-6094	323	11	pritzel	pritzel	NOUN
cana-6094	323	12	,	,	PUNCT
cana-6094	323	13	a.	a.	NOUN
cana-6094	323	14	,	,	PUNCT
cana-6094	323	15	green	green	ADJ
cana-6094	323	16	,	,	PUNCT
cana-6094	323	17	t.	t.	PROPN
cana-6094	323	18	,	,	PUNCT
cana-6094	323	19	figurnov	figurnov	PROPN
cana-6094	323	20	,	,	PUNCT
cana-6094	323	21	m.	m.	NOUN
cana-6094	323	22	,	,	PUNCT
cana-6094	323	23	ronneberger	ronneberger	NOUN
cana-6094	323	24	,	,	PUNCT
cana-6094	323	25	o.	o.	PROPN
cana-6094	323	26	,	,	PUNCT
cana-6094	323	27	...	...	PUNCT
cana-6094	323	28	&	&	CCONJ
cana-6094	323	29	hassabis	hassabis	PROPN
cana-6094	323	30	,	,	PUNCT
cana-6094	323	31	d.	d.	PROPN
cana-6094	323	32	(	(	PUNCT
cana-6094	323	33	2021	2021	NUM
cana-6094	323	34	)	)	PUNCT
cana-6094	323	35	.	.	PUNCT
cana-6094	324	1	highly	highly	ADV
cana-6094	324	2	accurate	accurate	ADJ
cana-6094	324	3	protein	protein	NOUN
cana-6094	324	4	structure	structure	NOUN
cana-6094	324	5	prediction	prediction	NOUN
cana-6094	324	6	with	with	ADP
cana-6094	324	7	alphafold2	alphafold2	NOUN
cana-6094	324	8	.	.	PUNCT
cana-6094	325	1	nature	nature	NOUN
cana-6094	325	2	.	.	PUNCT
cana-6094	326	1	48	48	NUM
cana-6094	326	2	.	.	PUNCT
cana-6094	326	3	sanchez	sanchez	PROPN
cana-6094	326	4	-	-	PUNCT
cana-6094	326	5	gonzalez	gonzalez	PROPN
cana-6094	326	6	,	,	PUNCT
cana-6094	326	7	a.	a.	PROPN
cana-6094	326	8	,	,	PUNCT
cana-6094	326	9	godwin	godwin	PROPN
cana-6094	326	10	,	,	PUNCT
cana-6094	326	11	j.	j.	PROPN
cana-6094	326	12	,	,	PUNCT
cana-6094	326	13	pfaff	pfaff	PROPN
cana-6094	326	14	,	,	PUNCT
cana-6094	326	15	t.	t.	PROPN
cana-6094	326	16	,	,	PUNCT
cana-6094	326	17	ying	ying	PROPN
cana-6094	326	18	,	,	PUNCT
cana-6094	326	19	r.	r.	PROPN
cana-6094	326	20	,	,	PUNCT
cana-6094	326	21	leskovec	leskovec	PROPN
cana-6094	326	22	,	,	PUNCT
cana-6094	326	23	j.	j.	PROPN
cana-6094	326	24	,	,	PUNCT
cana-6094	326	25	&	&	CCONJ
cana-6094	326	26	battaglia	battaglia	PROPN
cana-6094	326	27	,	,	PUNCT
cana-6094	326	28	p.	p.	NOUN
cana-6094	326	29	w.	w.	PROPN
cana-6094	326	30	(	(	PUNCT
cana-6094	326	31	2020	2020	NUM
cana-6094	326	32	)	)	PUNCT
cana-6094	326	33	.	.	PUNCT
cana-6094	327	1	learning	learn	VERB
cana-6094	327	2	to	to	PART
cana-6094	327	3	simulate	simulate	VERB
cana-6094	327	4	complex	complex	ADJ
cana-6094	327	5	physics	physics	NOUN
cana-6094	327	6	with	with	ADP
cana-6094	327	7	graph	graph	NOUN
cana-6094	327	8	networks	network	NOUN
cana-6094	327	9	.	.	PUNCT
cana-6094	328	1	international	international	ADJ
cana-6094	328	2	conference	conference	NOUN
cana-6094	328	3	on	on	ADP
cana-6094	328	4	machine	machine	NOUN
cana-6094	328	5	learning	learning	NOUN
cana-6094	328	6	(	(	PUNCT
cana-6094	328	7	icml	icml	PROPN
cana-6094	328	8	)	)	PUNCT
cana-6094	328	9	.	.	PUNCT
cana-6094	329	1	49	49	NUM
cana-6094	329	2	.	.	PUNCT
cana-6094	330	1	fey	fey	PROPN
cana-6094	330	2	,	,	PUNCT
cana-6094	330	3	m.	m.	NOUN
cana-6094	330	4	,	,	PUNCT
cana-6094	330	5	&	&	CCONJ
cana-6094	330	6	lenssen	lenssen	PROPN
cana-6094	330	7	,	,	PUNCT
cana-6094	330	8	j.	j.	PROPN
cana-6094	330	9	e.	e.	PROPN
cana-6094	330	10	(	(	PUNCT
cana-6094	330	11	2019	2019	NUM
cana-6094	330	12	)	)	PUNCT
cana-6094	330	13	.	.	PUNCT
cana-6094	331	1	fast	fast	ADJ
cana-6094	331	2	graph	graph	NOUN
cana-6094	331	3	representation	representation	NOUN
cana-6094	331	4	learning	learn	VERB
cana-6094	331	5	with	with	ADP
cana-6094	331	6	pytorch	pytorch	NOUN
cana-6094	331	7	geometric	geometric	ADJ
cana-6094	331	8	.	.	PUNCT
cana-6094	332	1	iclr	iclr	ADJ
cana-6094	332	2	workshop	workshop	NOUN
cana-6094	332	3	on	on	ADP
cana-6094	332	4	representation	representation	NOUN
cana-6094	332	5	learning	learn	VERB
cana-6094	332	6	on	on	ADP
cana-6094	332	7	graphs	graph	NOUN
cana-6094	332	8	and	and	CCONJ
cana-6094	332	9	manifolds	manifold	NOUN
cana-6094	332	10	.	.	PUNCT
cana-6094	333	1	53	53	NUM
cana-6094	333	2	.	.	PUNCT
cana-6094	334	1	wang	wang	PROPN
cana-6094	334	2	,	,	PUNCT
cana-6094	334	3	h.	h.	PROPN
cana-6094	334	4	,	,	PUNCT
cana-6094	334	5	&	&	CCONJ
cana-6094	334	6	zhang	zhang	PROPN
cana-6094	334	7	,	,	PUNCT
cana-6094	334	8	m.	m.	NOUN
cana-6094	334	9	(	(	PUNCT
cana-6094	334	10	2022	2022	NUM
cana-6094	334	11	)	)	PUNCT
cana-6094	334	12	.	.	PUNCT
cana-6094	335	1	graph	graph	NOUN
cana-6094	335	2	learning	learning	NOUN
cana-6094	335	3	:	:	PUNCT
cana-6094	335	4	a	a	DET
cana-6094	335	5	survey	survey	NOUN
cana-6094	335	6	.	.	PUNCT
cana-6094	336	1	ieee	ieee	NOUN
cana-6094	336	2	transactions	transaction	NOUN
cana-6094	336	3	on	on	ADP
cana-6094	336	4	artificial	artificial	ADJ
cana-6094	336	5	intelligence	intelligence	NOUN
cana-6094	336	6	.	.	PUNCT
cana-6094	337	1	50	50	NUM
cana-6094	337	2	.	.	X
cana-6094	338	1	topping	top	VERB
cana-6094	338	2	,	,	PUNCT
cana-6094	338	3	j.	j.	PROPN
cana-6094	338	4	,	,	PUNCT
cana-6094	338	5	di	di	X
cana-6094	338	6	giovanni	giovanni	PROPN
cana-6094	338	7	,	,	PUNCT
cana-6094	338	8	f.	f.	PROPN
cana-6094	338	9	,	,	PUNCT
cana-6094	338	10	chamberlain	chamberlain	PROPN
cana-6094	338	11	,	,	PUNCT
cana-6094	338	12	b.	b.	PROPN
cana-6094	338	13	p.	p.	PROPN
cana-6094	338	14	,	,	PUNCT
cana-6094	338	15	dong	dong	PROPN
cana-6094	338	16	,	,	PUNCT
cana-6094	338	17	x.	x.	PROPN
cana-6094	338	18	,	,	PUNCT
cana-6094	338	19	&	&	CCONJ
cana-6094	338	20	bronstein	bronstein	PROPN
cana-6094	338	21	,	,	PUNCT
cana-6094	338	22	m.	m.	NOUN
cana-6094	338	23	m.	m.	NOUN
cana-6094	338	24	(	(	PUNCT
cana-6094	338	25	2022	2022	NUM
cana-6094	338	26	)	)	PUNCT
cana-6094	338	27	.	.	PUNCT
cana-6094	339	1	understanding	understand	VERB
cana-6094	339	2	oversquashing	oversquashe	VERB
cana-6094	339	3	and	and	CCONJ
cana-6094	339	4	bottlenecks	bottleneck	NOUN
cana-6094	339	5	on	on	ADP
cana-6094	339	6	graphs	graph	NOUN
cana-6094	339	7	via	via	ADP
cana-6094	339	8	curvature	curvature	NOUN
cana-6094	339	9	.	.	PUNCT
cana-6094	340	1	international	international	ADJ
cana-6094	340	2	conference	conference	NOUN
cana-6094	340	3	on	on	ADP
cana-6094	340	4	learning	learn	VERB
cana-6094	340	5	representations	representation	NOUN
cana-6094	340	6	(	(	PUNCT
cana-6094	340	7	iclr	iclr	NOUN
cana-6094	340	8	)	)	PUNCT
cana-6094	340	9	.	.	PUNCT
cana-6094	341	1	51	51	NUM
cana-6094	341	2	.	.	X
cana-6094	341	3	beaini	beaini	PROPN
cana-6094	341	4	,	,	PUNCT
cana-6094	341	5	d.	d.	PROPN
cana-6094	341	6	,	,	PUNCT
cana-6094	341	7	passaro	passaro	NOUN
cana-6094	341	8	,	,	PUNCT
cana-6094	341	9	s.	s.	PROPN
cana-6094	341	10	,	,	PUNCT
cana-6094	341	11	létourneau	létourneau	PROPN
cana-6094	341	12	,	,	PUNCT
cana-6094	341	13	v.	v.	PROPN
cana-6094	341	14	,	,	PUNCT
cana-6094	341	15	hamilton	hamilton	PROPN
cana-6094	341	16	,	,	PUNCT
cana-6094	341	17	w.	w.	PROPN
cana-6094	341	18	l.	l.	PROPN
cana-6094	341	19	,	,	PUNCT
cana-6094	341	20	corso	corso	PROPN
cana-6094	341	21	,	,	PUNCT
cana-6094	341	22	g.	g.	PROPN
cana-6094	341	23	,	,	PUNCT
cana-6094	341	24	&	&	CCONJ
cana-6094	341	25	liò	liò	PROPN
cana-6094	341	26	,	,	PUNCT
cana-6094	341	27	p.	p.	NOUN
cana-6094	341	28	(	(	PUNCT
cana-6094	341	29	2021	2021	NUM
cana-6094	341	30	)	)	PUNCT
cana-6094	341	31	.	.	PUNCT
cana-6094	342	1	directional	directional	ADJ
cana-6094	342	2	graph	graph	NOUN
cana-6094	342	3	networks	network	NOUN
cana-6094	342	4	.	.	PUNCT
cana-6094	343	1	international	international	ADJ
cana-6094	343	2	conference	conference	NOUN
cana-6094	343	3	on	on	ADP
cana-6094	343	4	machine	machine	NOUN
cana-6094	343	5	learning	learning	NOUN
cana-6094	343	6	(	(	PUNCT
cana-6094	343	7	icml	icml	PROPN
cana-6094	343	8	)	)	PUNCT
cana-6094	343	9	.	.	PUNCT
cana-6094	344	1	52	52	NUM
cana-6094	344	2	.	.	PUNCT
cana-6094	344	3	kreuzer	kreuzer	PROPN
cana-6094	344	4	,	,	PUNCT
cana-6094	344	5	d.	d.	PROPN
cana-6094	344	6	,	,	PUNCT
cana-6094	344	7	beaini	beaini	PROPN
cana-6094	344	8	,	,	PUNCT
cana-6094	344	9	d.	d.	PROPN
cana-6094	344	10	,	,	PUNCT
cana-6094	344	11	hamilton	hamilton	PROPN
cana-6094	344	12	,	,	PUNCT
cana-6094	344	13	w.	w.	PROPN
cana-6094	344	14	l.	l.	PROPN
cana-6094	344	15	,	,	PUNCT
cana-6094	344	16	létourneau	létourneau	PROPN
cana-6094	344	17	,	,	PUNCT
cana-6094	344	18	v.	v.	ADV
cana-6094	344	19	,	,	PUNCT
cana-6094	344	20	&	&	CCONJ
cana-6094	344	21	tossou	tossou	NOUN
cana-6094	344	22	,	,	PUNCT
cana-6094	344	23	p.	p.	NOUN
cana-6094	344	24	(	(	PUNCT
cana-6094	344	25	2021	2021	NUM
cana-6094	344	26	)	)	PUNCT
cana-6094	344	27	.	.	PUNCT
cana-6094	345	1	rethinking	rethink	VERB
cana-6094	345	2	graph	graph	NOUN
cana-6094	345	3	transformers	transformer	NOUN
cana-6094	345	4	with	with	ADP
cana-6094	345	5	spectral	spectral	ADJ
cana-6094	345	6	attention	attention	NOUN
cana-6094	345	7	.	.	PUNCT
cana-6094	346	1	advances	advance	NOUN
cana-6094	346	2	in	in	ADP
cana-6094	346	3	neural	neural	ADJ
cana-6094	346	4	information	information	NOUN
cana-6094	346	5	processing	processing	NOUN
cana-6094	346	6	systems	system	NOUN
cana-6094	346	7	(	(	PUNCT
cana-6094	346	8	neurips	neurip	NOUN
cana-6094	346	9	)	)	PUNCT
cana-6094	346	10	.	.	PUNCT
cana-6094	347	1	53	53	NUM
cana-6094	347	2	.	.	PUNCT
cana-6094	347	3	hu	hu	PROPN
cana-6094	347	4	,	,	PUNCT
cana-6094	347	5	w.	w.	PROPN
cana-6094	347	6	,	,	PUNCT
cana-6094	347	7	fey	fey	PROPN
cana-6094	347	8	,	,	PUNCT
cana-6094	347	9	m.	m.	NOUN
cana-6094	347	10	,	,	PUNCT
cana-6094	347	11	zitnik	zitnik	NOUN
cana-6094	347	12	,	,	PUNCT
cana-6094	347	13	m.	m.	NOUN
cana-6094	347	14	,	,	PUNCT
cana-6094	347	15	dong	dong	PROPN
cana-6094	347	16	,	,	PUNCT
cana-6094	347	17	y.	y.	PROPN
cana-6094	347	18	,	,	PUNCT
cana-6094	347	19	ren	ren	PROPN
cana-6094	347	20	,	,	PUNCT
cana-6094	347	21	h.	h.	PROPN
cana-6094	347	22	,	,	PUNCT
cana-6094	347	23	liu	liu	PROPN
cana-6094	347	24	,	,	PUNCT
cana-6094	347	25	b.	b.	PROPN
cana-6094	347	26	,	,	PUNCT
cana-6094	347	27	...	...	PUNCT
cana-6094	347	28	&	&	CCONJ
cana-6094	347	29	leskovec	leskovec	PROPN
cana-6094	347	30	,	,	PUNCT
cana-6094	347	31	j.	j.	PROPN
cana-6094	347	32	(	(	PUNCT
cana-6094	347	33	2021	2021	NUM
cana-6094	347	34	)	)	PUNCT
cana-6094	347	35	.	.	PUNCT
cana-6094	348	1	open	open	ADJ
cana-6094	348	2	graph	graph	NOUN
cana-6094	348	3	benchmark	benchmark	NOUN
cana-6094	348	4	:	:	PUNCT
cana-6094	348	5	datasets	dataset	NOUN
cana-6094	348	6	for	for	ADP
cana-6094	348	7	machine	machine	NOUN
cana-6094	348	8	learning	learn	VERB
cana-6094	348	9	on	on	ADP
cana-6094	348	10	graphs	graph	NOUN
cana-6094	348	11	.	.	PUNCT
cana-6094	349	1	advances	advance	NOUN
cana-6094	349	2	in	in	ADP
cana-6094	349	3	neural	neural	ADJ
cana-6094	349	4	information	information	NOUN
cana-6094	349	5	processing	processing	NOUN
cana-6094	349	6	systems	system	NOUN
cana-6094	349	7	(	(	PUNCT
cana-6094	349	8	neurips	neurip	NOUN
cana-6094	349	9	)	)	PUNCT
cana-6094	349	10	.	.	PUNCT
cana-6094	350	1	54	54	NUM
cana-6094	350	2	.	.	X
cana-6094	351	1	wang	wang	PROPN
cana-6094	351	2	,	,	PUNCT
cana-6094	351	3	x.	x.	PROPN
cana-6094	351	4	,	,	PUNCT
cana-6094	351	5	cui	cui	PROPN
cana-6094	351	6	,	,	PUNCT
cana-6094	351	7	p.	p.	PROPN
cana-6094	351	8	,	,	PUNCT
cana-6094	351	9	wang	wang	PROPN
cana-6094	351	10	,	,	PUNCT
cana-6094	351	11	j.	j.	PROPN
cana-6094	351	12	,	,	PUNCT
cana-6094	351	13	pei	pei	PROPN
cana-6094	351	14	,	,	PUNCT
cana-6094	351	15	j.	j.	PROPN
cana-6094	351	16	,	,	PUNCT
cana-6094	351	17	zhu	zhu	PROPN
cana-6094	351	18	,	,	PUNCT
cana-6094	351	19	w.	w.	PROPN
cana-6094	351	20	,	,	PUNCT
cana-6094	351	21	&	&	CCONJ
cana-6094	351	22	yang	yang	PROPN
cana-6094	351	23	,	,	PUNCT
cana-6094	351	24	s.	s.	PROPN
cana-6094	351	25	(	(	PUNCT
cana-6094	351	26	2017	2017	NUM
cana-6094	351	27	)	)	PUNCT
cana-6094	351	28	.	.	PUNCT
cana-6094	352	1	community	community	NOUN
cana-6094	352	2	preserving	preserve	VERB
cana-6094	352	3	network	network	NOUN
cana-6094	352	4	embedding	embed	VERB
cana-6094	352	5	.	.	PUNCT
cana-6094	353	1	proceedings	proceeding	NOUN
cana-6094	353	2	of	of	ADP
cana-6094	353	3	the	the	DET
cana-6094	353	4	aaai	aaai	PROPN
cana-6094	353	5	conference	conference	NOUN
cana-6094	353	6	on	on	ADP
cana-6094	353	7	artificial	artificial	ADJ
cana-6094	353	8	intelligence	intelligence	NOUN
cana-6094	353	9	.	.	PUNCT