id sid tid token lemma pos ajst-6287 1 1 academic academic ADJ ajst-6287 1 2 journal journal NOUN ajst-6287 1 3 of of ADP ajst-6287 1 4 science science NOUN ajst-6287 1 5 and and CCONJ ajst-6287 1 6 technology technology NOUN ajst-6287 1 7 issn issn NOUN ajst-6287 1 8 : : PUNCT ajst-6287 1 9 2771 2771 NUM ajst-6287 1 10 - - SYM ajst-6287 1 11 3032 3032 NUM ajst-6287 1 12 | | NOUN ajst-6287 1 13 vol vol NOUN ajst-6287 1 14 . . PROPN ajst-6287 2 1 5 5 NUM ajst-6287 2 2 , , PUNCT ajst-6287 2 3 no no INTJ ajst-6287 2 4 . . NOUN ajst-6287 2 5 2 2 NUM ajst-6287 2 6 , , PUNCT ajst-6287 2 7 2023 2023 NUM ajst-6287 2 8 82 82 NUM ajst-6287 2 9 plant plant NOUN ajst-6287 2 10 leaf leaf NOUN ajst-6287 2 11 diseases disease NOUN ajst-6287 2 12 classification classification NOUN ajst-6287 2 13 based base VERB ajst-6287 2 14 on on ADP ajst-6287 2 15 few‐shot few‐shot PROPN ajst-6287 2 16 learning learning PROPN ajst-6287 2 17 wei wei PROPN ajst-6287 2 18 wu wu PROPN ajst-6287 2 19 taishan taishan PROPN ajst-6287 2 20 university university PROPN ajst-6287 2 21 , , PUNCT ajst-6287 2 22 tai'an tai'an PROPN ajst-6287 2 23 , , PUNCT ajst-6287 2 24 271000 271000 NUM ajst-6287 2 25 , , PUNCT ajst-6287 2 26 china china PROPN ajst-6287 2 27 abstract abstract NOUN ajst-6287 2 28 : : PUNCT ajst-6287 2 29 plant plant NOUN ajst-6287 2 30 leaf leaf NOUN ajst-6287 2 31 disease disease NOUN ajst-6287 2 32 is be AUX ajst-6287 2 33 one one NUM ajst-6287 2 34 of of ADP ajst-6287 2 35 the the DET ajst-6287 2 36 important important ADJ ajst-6287 2 37 factors factor NOUN ajst-6287 2 38 affecting affect VERB ajst-6287 2 39 the the DET ajst-6287 2 40 normal normal ADJ ajst-6287 2 41 growth growth NOUN ajst-6287 2 42 of of ADP ajst-6287 2 43 plants plant NOUN ajst-6287 2 44 . . PUNCT ajst-6287 3 1 it it PRON ajst-6287 3 2 is be AUX ajst-6287 3 3 important important ADJ ajst-6287 3 4 to to PART ajst-6287 3 5 accurately accurately ADV ajst-6287 3 6 identify identify VERB ajst-6287 3 7 the the DET ajst-6287 3 8 types type NOUN ajst-6287 3 9 of of ADP ajst-6287 3 10 leaf leaf NOUN ajst-6287 3 11 disease disease NOUN ajst-6287 3 12 and and CCONJ ajst-6287 3 13 take take VERB ajst-6287 3 14 effective effective ADJ ajst-6287 3 15 measures measure NOUN ajst-6287 3 16 to to PART ajst-6287 3 17 ensure ensure VERB ajst-6287 3 18 the the DET ajst-6287 3 19 increase increase NOUN ajst-6287 3 20 of of ADP ajst-6287 3 21 crop crop NOUN ajst-6287 3 22 production production NOUN ajst-6287 3 23 . . PUNCT ajst-6287 4 1 in in ADP ajst-6287 4 2 this this DET ajst-6287 4 3 study study NOUN ajst-6287 4 4 , , PUNCT ajst-6287 4 5 we we PRON ajst-6287 4 6 propose propose VERB ajst-6287 4 7 a a DET ajst-6287 4 8 plant plant NOUN ajst-6287 4 9 leaf leaf NOUN ajst-6287 4 10 diseases disease NOUN ajst-6287 4 11 classification classification NOUN ajst-6287 4 12 method method NOUN ajst-6287 4 13 based base VERB ajst-6287 4 14 on on ADP ajst-6287 4 15 few few ADJ ajst-6287 4 16 - - PUNCT ajst-6287 4 17 shot shot NOUN ajst-6287 4 18 learning learning NOUN ajst-6287 4 19 ( ( PUNCT ajst-6287 4 20 fsl fsl PROPN ajst-6287 4 21 ) ) PUNCT ajst-6287 4 22 , , PUNCT ajst-6287 4 23 which which PRON ajst-6287 4 24 can can AUX ajst-6287 4 25 obtain obtain VERB ajst-6287 4 26 good good ADJ ajst-6287 4 27 identification identification NOUN ajst-6287 4 28 accuracy accuracy NOUN ajst-6287 4 29 in in ADP ajst-6287 4 30 the the DET ajst-6287 4 31 classification classification NOUN ajst-6287 4 32 task task NOUN ajst-6287 4 33 with with ADP ajst-6287 4 34 few few ADJ ajst-6287 4 35 samples sample NOUN ajst-6287 4 36 . . PUNCT ajst-6287 5 1 by by ADP ajst-6287 5 2 changing change VERB ajst-6287 5 3 the the DET ajst-6287 5 4 parameter parameter NOUN ajst-6287 5 5 setting setting NOUN ajst-6287 5 6 in in ADP ajst-6287 5 7 fsl fsl PROPN ajst-6287 5 8 , , PUNCT ajst-6287 5 9 some some DET ajst-6287 5 10 key key ADJ ajst-6287 5 11 features feature NOUN ajst-6287 5 12 affecting affect VERB ajst-6287 5 13 classification classification NOUN ajst-6287 5 14 accuracy accuracy NOUN ajst-6287 5 15 are be AUX ajst-6287 5 16 revealed reveal VERB ajst-6287 5 17 . . PUNCT ajst-6287 6 1 the the DET ajst-6287 6 2 proposal proposal NOUN ajst-6287 6 3 of of ADP ajst-6287 6 4 this this DET ajst-6287 6 5 research research NOUN ajst-6287 6 6 method method NOUN ajst-6287 6 7 expands expand VERB ajst-6287 6 8 the the DET ajst-6287 6 9 application application NOUN ajst-6287 6 10 scope scope NOUN ajst-6287 6 11 of of ADP ajst-6287 6 12 artificial artificial ADJ ajst-6287 6 13 intelligence intelligence NOUN ajst-6287 6 14 in in ADP ajst-6287 6 15 agriculture agriculture NOUN ajst-6287 6 16 , , PUNCT ajst-6287 6 17 which which PRON ajst-6287 6 18 has have VERB ajst-6287 6 19 important important ADJ ajst-6287 6 20 practical practical ADJ ajst-6287 6 21 significance significance NOUN ajst-6287 6 22 . . PUNCT ajst-6287 7 1 keywords keyword NOUN ajst-6287 7 2 : : PUNCT ajst-6287 7 3 convolutional convolutional ADJ ajst-6287 7 4 neural neural ADJ ajst-6287 7 5 network network NOUN ajst-6287 7 6 , , PUNCT ajst-6287 7 7 few few ADJ ajst-6287 7 8 - - PUNCT ajst-6287 7 9 shot shot NOUN ajst-6287 7 10 learning learning NOUN ajst-6287 7 11 , , PUNCT ajst-6287 7 12 meta meta PROPN ajst-6287 7 13 learning learning NOUN ajst-6287 7 14 , , PUNCT ajst-6287 7 15 plant plant NOUN ajst-6287 7 16 leaf leaf NOUN ajst-6287 7 17 disease disease NOUN ajst-6287 7 18 identification identification NOUN ajst-6287 7 19 . . PUNCT ajst-6287 8 1 1 1 X ajst-6287 8 2 . . X ajst-6287 8 3 introduction introduction NOUN ajst-6287 8 4 in in ADP ajst-6287 8 5 modern modern ADJ ajst-6287 8 6 agriculture agriculture NOUN ajst-6287 8 7 , , PUNCT ajst-6287 8 8 plant plant NOUN ajst-6287 8 9 leaf leaf NOUN ajst-6287 8 10 disease disease NOUN ajst-6287 8 11 is be AUX ajst-6287 8 12 one one NUM ajst-6287 8 13 of of ADP ajst-6287 8 14 the the DET ajst-6287 8 15 important important ADJ ajst-6287 8 16 factors factor NOUN ajst-6287 8 17 affecting affect VERB ajst-6287 8 18 the the DET ajst-6287 8 19 normal normal ADJ ajst-6287 8 20 plant plant NOUN ajst-6287 8 21 growth growth NOUN ajst-6287 8 22 . . PUNCT ajst-6287 9 1 it it PRON ajst-6287 9 2 is be AUX ajst-6287 9 3 of of ADP ajst-6287 9 4 great great ADJ ajst-6287 9 5 significance significance NOUN ajst-6287 9 6 to to PART ajst-6287 9 7 identify identify VERB ajst-6287 9 8 the the DET ajst-6287 9 9 types type NOUN ajst-6287 9 10 of of ADP ajst-6287 9 11 crop crop NOUN ajst-6287 9 12 leaf leaf NOUN ajst-6287 9 13 disease disease NOUN ajst-6287 9 14 timely timely ADV ajst-6287 9 15 and and CCONJ ajst-6287 9 16 accurately accurately ADV ajst-6287 9 17 and and CCONJ ajst-6287 9 18 to to PART ajst-6287 9 19 ensure ensure VERB ajst-6287 9 20 the the DET ajst-6287 9 21 steady steady ADJ ajst-6287 9 22 growth growth NOUN ajst-6287 9 23 of of ADP ajst-6287 9 24 crop crop NOUN ajst-6287 9 25 production production NOUN ajst-6287 9 26 and and CCONJ ajst-6287 9 27 quality quality NOUN ajst-6287 9 28 . . PUNCT ajst-6287 10 1 in in ADP ajst-6287 10 2 traditional traditional ADJ ajst-6287 10 3 ways way NOUN ajst-6287 10 4 , , PUNCT ajst-6287 10 5 it it PRON ajst-6287 10 6 ’s ’s AUX ajst-6287 10 7 mainly mainly ADV ajst-6287 10 8 relied rely VERB ajst-6287 10 9 on on ADP ajst-6287 10 10 artificial artificial ADJ ajst-6287 10 11 methods method NOUN ajst-6287 10 12 to to PART ajst-6287 10 13 identify identify VERB ajst-6287 10 14 plant plant NOUN ajst-6287 10 15 leaf leaf NOUN ajst-6287 10 16 disease disease NOUN ajst-6287 10 17 , , PUNCT ajst-6287 10 18 such such ADJ ajst-6287 10 19 as as ADP ajst-6287 10 20 consulting consult VERB ajst-6287 10 21 experts expert NOUN ajst-6287 10 22 or or CCONJ ajst-6287 10 23 experienced experienced ADJ ajst-6287 10 24 farmers farmer NOUN ajst-6287 10 25 . . PUNCT ajst-6287 11 1 however however ADV ajst-6287 11 2 , , PUNCT ajst-6287 11 3 the the DET ajst-6287 11 4 acquisition acquisition NOUN ajst-6287 11 5 of of ADP ajst-6287 11 6 these these DET ajst-6287 11 7 knowledge knowledge NOUN ajst-6287 11 8 has have VERB ajst-6287 11 9 some some DET ajst-6287 11 10 disadvantages disadvantage NOUN ajst-6287 11 11 , , PUNCT ajst-6287 11 12 such such ADJ ajst-6287 11 13 as as ADP ajst-6287 11 14 high high ADJ ajst-6287 11 15 cost cost NOUN ajst-6287 11 16 , , PUNCT ajst-6287 11 17 low low ADJ ajst-6287 11 18 time time NOUN ajst-6287 11 19 efficiency efficiency NOUN ajst-6287 11 20 and and CCONJ ajst-6287 11 21 unstable unstable ADJ ajst-6287 11 22 accuracy accuracy NOUN ajst-6287 11 23 . . PUNCT ajst-6287 12 1 with with ADP ajst-6287 12 2 the the DET ajst-6287 12 3 continuous continuous ADJ ajst-6287 12 4 development development NOUN ajst-6287 12 5 of of ADP ajst-6287 12 6 computer computer NOUN ajst-6287 12 7 technology technology NOUN ajst-6287 12 8 , , PUNCT ajst-6287 12 9 the the DET ajst-6287 12 10 automatic automatic ADJ ajst-6287 12 11 identification identification NOUN ajst-6287 12 12 of of ADP ajst-6287 12 13 plant plant NOUN ajst-6287 12 14 leaf leaf NOUN ajst-6287 12 15 disease disease NOUN ajst-6287 12 16 by by ADP ajst-6287 12 17 intelligent intelligent ADJ ajst-6287 12 18 machine machine NOUN ajst-6287 12 19 has have AUX ajst-6287 12 20 gained gain VERB ajst-6287 12 21 more more ADJ ajst-6287 12 22 and and CCONJ ajst-6287 12 23 more more ADJ ajst-6287 12 24 attention attention NOUN ajst-6287 12 25 in in ADP ajst-6287 12 26 the the DET ajst-6287 12 27 modern modern ADJ ajst-6287 12 28 agricultural agricultural ADJ ajst-6287 12 29 management management NOUN ajst-6287 12 30 . . PUNCT ajst-6287 13 1 in in ADP ajst-6287 13 2 the the DET ajst-6287 13 3 past past ADJ ajst-6287 13 4 few few ADJ ajst-6287 13 5 years year NOUN ajst-6287 13 6 , , PUNCT ajst-6287 13 7 the the DET ajst-6287 13 8 plant plant NOUN ajst-6287 13 9 leaf leaf NOUN ajst-6287 13 10 disease disease NOUN ajst-6287 13 11 was be AUX ajst-6287 13 12 mainly mainly ADV ajst-6287 13 13 identified identify VERB ajst-6287 13 14 through through ADP ajst-6287 13 15 traditional traditional ADJ ajst-6287 13 16 machine machine NOUN ajst-6287 13 17 learning learning NOUN ajst-6287 13 18 methods method NOUN ajst-6287 13 19 such such ADJ ajst-6287 13 20 as as ADP ajst-6287 13 21 support support NOUN ajst-6287 13 22 vector vector NOUN ajst-6287 13 23 machines machine NOUN ajst-6287 13 24 , , PUNCT ajst-6287 13 25 k k NOUN ajst-6287 13 26 - - PUNCT ajst-6287 13 27 means mean NOUN ajst-6287 13 28 , , PUNCT ajst-6287 13 29 and and CCONJ ajst-6287 13 30 bayesian bayesian NOUN ajst-6287 13 31 etc etc X ajst-6287 13 32 . . X ajst-6287 14 1 with with ADP ajst-6287 14 2 the the DET ajst-6287 14 3 rapid rapid ADJ ajst-6287 14 4 improvement improvement NOUN ajst-6287 14 5 of of ADP ajst-6287 14 6 computing compute VERB ajst-6287 14 7 power power NOUN ajst-6287 14 8 and and CCONJ ajst-6287 14 9 the the DET ajst-6287 14 10 continuous continuous ADJ ajst-6287 14 11 development development NOUN ajst-6287 14 12 of of ADP ajst-6287 14 13 big big ADJ ajst-6287 14 14 data datum NOUN ajst-6287 14 15 technology technology NOUN ajst-6287 14 16 , , PUNCT ajst-6287 14 17 more more ADJ ajst-6287 14 18 and and CCONJ ajst-6287 14 19 more more ADJ ajst-6287 14 20 researchers researcher NOUN ajst-6287 14 21 turn turn VERB ajst-6287 14 22 their their PRON ajst-6287 14 23 attention attention NOUN ajst-6287 14 24 to to ADP ajst-6287 14 25 the the DET ajst-6287 14 26 field field NOUN ajst-6287 14 27 of of ADP ajst-6287 14 28 deep deep ADJ ajst-6287 14 29 learning learning NOUN ajst-6287 14 30 . . PUNCT ajst-6287 15 1 for for ADP ajst-6287 15 2 example example NOUN ajst-6287 15 3 , , PUNCT ajst-6287 15 4 ferentinos ferentinos ADV ajst-6287 15 5 etc etc X ajst-6287 15 6 . . PUNCT ajst-6287 16 1 [ [ X ajst-6287 16 2 1 1 X ajst-6287 16 3 ] ] PUNCT ajst-6287 16 4 use use VERB ajst-6287 16 5 a a DET ajst-6287 16 6 deep deep ADJ ajst-6287 16 7 convolutional convolutional ADJ ajst-6287 16 8 neural neural ADJ ajst-6287 16 9 network network NOUN ajst-6287 16 10 for for ADP ajst-6287 16 11 the the DET ajst-6287 16 12 detection detection NOUN ajst-6287 16 13 and and CCONJ ajst-6287 16 14 diagnosis diagnosis NOUN ajst-6287 16 15 of of ADP ajst-6287 16 16 plant plant NOUN ajst-6287 16 17 leaf leaf NOUN ajst-6287 16 18 diseases disease NOUN ajst-6287 16 19 , , PUNCT ajst-6287 16 20 which which PRON ajst-6287 16 21 can can AUX ajst-6287 16 22 achieve achieve VERB ajst-6287 16 23 a a DET ajst-6287 16 24 recognition recognition NOUN ajst-6287 16 25 accuracy accuracy NOUN ajst-6287 16 26 of of ADP ajst-6287 16 27 99.53 99.53 NUM ajst-6287 16 28 % % NOUN ajst-6287 16 29 . . PUNCT ajst-6287 17 1 the the DET ajst-6287 17 2 data datum NOUN ajst-6287 17 3 for for ADP ajst-6287 17 4 model model NOUN ajst-6287 17 5 training training NOUN ajst-6287 17 6 are be AUX ajst-6287 17 7 obtained obtain VERB ajst-6287 17 8 from from ADP ajst-6287 17 9 a a DET ajst-6287 17 10 publicly publicly ADV ajst-6287 17 11 available available ADJ ajst-6287 17 12 database database NOUN ajst-6287 17 13 which which PRON ajst-6287 17 14 contains contain VERB ajst-6287 17 15 87,848 87,848 NUM ajst-6287 17 16 images image NOUN ajst-6287 17 17 . . PUNCT ajst-6287 18 1 zhou zhou PROPN ajst-6287 18 2 etc etc X ajst-6287 18 3 . . PUNCT ajst-6287 19 1 [ [ X ajst-6287 19 2 2 2 X ajst-6287 19 3 ] ] X ajst-6287 19 4 propose propose VERB ajst-6287 19 5 a a DET ajst-6287 19 6 restructured restructure VERB ajst-6287 19 7 residual residual ADJ ajst-6287 19 8 dense dense ADJ ajst-6287 19 9 network network NOUN ajst-6287 19 10 by by ADP ajst-6287 19 11 combining combine VERB ajst-6287 19 12 the the DET ajst-6287 19 13 advantages advantage NOUN ajst-6287 19 14 of of ADP ajst-6287 19 15 deep deep ADJ ajst-6287 19 16 residual residual ADJ ajst-6287 19 17 networks network NOUN ajst-6287 19 18 and and CCONJ ajst-6287 19 19 dense dense ADJ ajst-6287 19 20 networks network NOUN ajst-6287 19 21 for for ADP ajst-6287 19 22 tomato tomato NOUN ajst-6287 19 23 leaf leaf NOUN ajst-6287 19 24 disease disease NOUN ajst-6287 19 25 identification identification NOUN ajst-6287 19 26 . . PUNCT ajst-6287 20 1 it it PRON ajst-6287 20 2 can can AUX ajst-6287 20 3 obtain obtain VERB ajst-6287 20 4 significant significant ADJ ajst-6287 20 5 improvements improvement NOUN ajst-6287 20 6 over over ADP ajst-6287 20 7 most most ADJ ajst-6287 20 8 of of ADP ajst-6287 20 9 the the DET ajst-6287 20 10 state state NOUN ajst-6287 20 11 - - PUNCT ajst-6287 20 12 of of ADP ajst-6287 20 13 - - PUNCT ajst-6287 20 14 the the DET ajst-6287 20 15 - - PUNCT ajst-6287 20 16 art art NOUN ajst-6287 20 17 models model NOUN ajst-6287 20 18 in in ADP ajst-6287 20 19 crop crop NOUN ajst-6287 20 20 leaf leaf NOUN ajst-6287 20 21 identification identification NOUN ajst-6287 20 22 . . PUNCT ajst-6287 21 1 deep deep ADJ ajst-6287 21 2 learning learning NOUN ajst-6287 21 3 is be AUX ajst-6287 21 4 one one NUM ajst-6287 21 5 of of ADP ajst-6287 21 6 the the DET ajst-6287 21 7 important important ADJ ajst-6287 21 8 branches branch NOUN ajst-6287 21 9 in in ADP ajst-6287 21 10 the the DET ajst-6287 21 11 field field NOUN ajst-6287 21 12 of of ADP ajst-6287 21 13 machine machine NOUN ajst-6287 21 14 learning learning NOUN ajst-6287 21 15 . . PUNCT ajst-6287 22 1 its its PRON ajst-6287 22 2 deep deep ADJ ajst-6287 22 3 network network NOUN ajst-6287 22 4 structure structure NOUN ajst-6287 22 5 corresponds correspond VERB ajst-6287 22 6 to to ADP ajst-6287 22 7 the the DET ajst-6287 22 8 huge huge ADJ ajst-6287 22 9 model model NOUN ajst-6287 22 10 parameters parameter NOUN ajst-6287 22 11 . . PUNCT ajst-6287 23 1 a a DET ajst-6287 23 2 large large ADJ ajst-6287 23 3 amount amount NOUN ajst-6287 23 4 of of ADP ajst-6287 23 5 sample sample NOUN ajst-6287 23 6 data datum NOUN ajst-6287 23 7 is be AUX ajst-6287 23 8 usually usually ADV ajst-6287 23 9 required require VERB ajst-6287 23 10 to to PART ajst-6287 23 11 train train VERB ajst-6287 23 12 these these DET ajst-6287 23 13 parameters parameter NOUN ajst-6287 23 14 to to PART ajst-6287 23 15 make make VERB ajst-6287 23 16 the the DET ajst-6287 23 17 model model NOUN ajst-6287 23 18 achieve achieve VERB ajst-6287 23 19 better well ADJ ajst-6287 23 20 results result NOUN ajst-6287 23 21 . . PUNCT ajst-6287 24 1 the the DET ajst-6287 24 2 network network NOUN ajst-6287 24 3 model model NOUN ajst-6287 24 4 is be AUX ajst-6287 24 5 easy easy ADJ ajst-6287 24 6 to to PART ajst-6287 24 7 overfit overfit VERB ajst-6287 24 8 if if SCONJ ajst-6287 24 9 there there PRON ajst-6287 24 10 are be VERB ajst-6287 24 11 few few ADJ ajst-6287 24 12 training training NOUN ajst-6287 24 13 data datum NOUN ajst-6287 24 14 . . PUNCT ajst-6287 25 1 in in ADP ajst-6287 25 2 many many ADJ ajst-6287 25 3 real real ADJ ajst-6287 25 4 - - PUNCT ajst-6287 25 5 world world NOUN ajst-6287 25 6 applications application NOUN ajst-6287 25 7 the the DET ajst-6287 25 8 acquisition acquisition NOUN ajst-6287 25 9 of of ADP ajst-6287 25 10 sample sample NOUN ajst-6287 25 11 data datum NOUN ajst-6287 25 12 is be AUX ajst-6287 25 13 very very ADV ajst-6287 25 14 difficult difficult ADJ ajst-6287 25 15 . . PUNCT ajst-6287 26 1 some some PRON ajst-6287 26 2 are be AUX ajst-6287 26 3 because because SCONJ ajst-6287 26 4 of of ADP ajst-6287 26 5 the the DET ajst-6287 26 6 high high ADJ ajst-6287 26 7 cost cost NOUN ajst-6287 26 8 of of ADP ajst-6287 26 9 human human ADJ ajst-6287 26 10 and and CCONJ ajst-6287 26 11 material material ADJ ajst-6287 26 12 resources resource NOUN ajst-6287 26 13 , , PUNCT ajst-6287 26 14 while while SCONJ ajst-6287 26 15 some some PRON ajst-6287 26 16 are be AUX ajst-6287 26 17 small small ADJ ajst-6287 26 18 probability probability NOUN ajst-6287 26 19 events event NOUN ajst-6287 26 20 and and CCONJ ajst-6287 26 21 can can AUX ajst-6287 26 22 not not PART ajst-6287 26 23 be be AUX ajst-6287 26 24 collected collect VERB ajst-6287 26 25 in in ADP ajst-6287 26 26 large large ADJ ajst-6287 26 27 quantities quantity NOUN ajst-6287 26 28 in in ADP ajst-6287 26 29 the the DET ajst-6287 26 30 short short ADJ ajst-6287 26 31 term term NOUN ajst-6287 26 32 . . PUNCT ajst-6287 27 1 on on ADP ajst-6287 27 2 the the DET ajst-6287 27 3 contrary contrary NOUN ajst-6287 27 4 , , PUNCT ajst-6287 27 5 it it PRON ajst-6287 27 6 does do AUX ajst-6287 27 7 not not PART ajst-6287 27 8 need need VERB ajst-6287 27 9 a a DET ajst-6287 27 10 lot lot NOUN ajst-6287 27 11 of of ADP ajst-6287 27 12 data datum NOUN ajst-6287 27 13 to to PART ajst-6287 27 14 learn learn VERB ajst-6287 27 15 new new ADJ ajst-6287 27 16 things thing NOUN ajst-6287 27 17 for for ADP ajst-6287 27 18 human human ADJ ajst-6287 27 19 being being NOUN ajst-6287 27 20 . . PUNCT ajst-6287 28 1 for for ADP ajst-6287 28 2 example example NOUN ajst-6287 28 3 , , PUNCT ajst-6287 28 4 infants infant NOUN ajst-6287 28 5 are be AUX ajst-6287 28 6 often often ADV ajst-6287 28 7 taught teach VERB ajst-6287 28 8 a a DET ajst-6287 28 9 new new ADJ ajst-6287 28 10 animal animal NOUN ajst-6287 28 11 by by ADP ajst-6287 28 12 only only ADV ajst-6287 28 13 a a DET ajst-6287 28 14 single single ADJ ajst-6287 28 15 picture picture NOUN ajst-6287 28 16 . . PUNCT ajst-6287 29 1 in in ADP ajst-6287 29 2 order order NOUN ajst-6287 29 3 to to PART ajst-6287 29 4 enable enable VERB ajst-6287 29 5 computers computer NOUN ajst-6287 29 6 to to PART ajst-6287 29 7 learn learn VERB ajst-6287 29 8 like like ADP ajst-6287 29 9 humans human NOUN ajst-6287 29 10 , , PUNCT ajst-6287 29 11 a a DET ajst-6287 29 12 new new ADJ ajst-6287 29 13 branch branch NOUN ajst-6287 29 14 of of ADP ajst-6287 29 15 deep deep ADJ ajst-6287 29 16 learning learning NOUN ajst-6287 29 17 called call VERB ajst-6287 29 18 few few ADJ ajst-6287 29 19 - - PUNCT ajst-6287 29 20 shot shot NOUN ajst-6287 29 21 learning learning NOUN ajst-6287 29 22 ( ( PUNCT ajst-6287 29 23 fsl fsl PROPN ajst-6287 29 24 ) ) PUNCT ajst-6287 29 25 , , PUNCT ajst-6287 29 26 has have AUX ajst-6287 29 27 gradually gradually ADV ajst-6287 29 28 aroused arouse VERB ajst-6287 29 29 the the DET ajst-6287 29 30 interest interest NOUN ajst-6287 29 31 of of ADP ajst-6287 29 32 more more ADJ ajst-6287 29 33 and and CCONJ ajst-6287 29 34 more more ADJ ajst-6287 29 35 researchers researcher NOUN ajst-6287 29 36 . . PUNCT ajst-6287 30 1 using use VERB ajst-6287 30 2 fsl fsl PROPN ajst-6287 30 3 to to PART ajst-6287 30 4 solve solve VERB ajst-6287 30 5 the the DET ajst-6287 30 6 classification classification NOUN ajst-6287 30 7 problem problem NOUN ajst-6287 30 8 means mean VERB ajst-6287 30 9 that that SCONJ ajst-6287 30 10 the the DET ajst-6287 30 11 model model NOUN ajst-6287 30 12 can can AUX ajst-6287 30 13 obtain obtain VERB ajst-6287 30 14 the the DET ajst-6287 30 15 recognition recognition NOUN ajst-6287 30 16 ability ability NOUN ajst-6287 30 17 of of ADP ajst-6287 30 18 new new ADJ ajst-6287 30 19 categories category NOUN ajst-6287 30 20 with with ADP ajst-6287 30 21 a a DET ajst-6287 30 22 small small ADJ ajst-6287 30 23 amount amount NOUN ajst-6287 30 24 of of ADP ajst-6287 30 25 training training NOUN ajst-6287 30 26 data datum NOUN ajst-6287 30 27 . . PUNCT ajst-6287 31 1 at at ADP ajst-6287 31 2 present present ADJ ajst-6287 31 3 , , PUNCT ajst-6287 31 4 there there PRON ajst-6287 31 5 are be VERB ajst-6287 31 6 three three NUM ajst-6287 31 7 main main ADJ ajst-6287 31 8 methods method NOUN ajst-6287 31 9 : : PUNCT ajst-6287 31 10 data datum NOUN ajst-6287 31 11 enhancement enhancement NOUN ajst-6287 31 12 , , PUNCT ajst-6287 31 13 transfer transfer NOUN ajst-6287 31 14 learning learning NOUN ajst-6287 31 15 and and CCONJ ajst-6287 31 16 meta meta NOUN ajst-6287 31 17 - - PUNCT ajst-6287 31 18 learning learning NOUN ajst-6287 31 19 . . PUNCT ajst-6287 32 1 the the DET ajst-6287 32 2 data data NOUN ajst-6287 32 3 enhancement enhancement NOUN ajst-6287 32 4 method method NOUN ajst-6287 32 5 mainly mainly ADV ajst-6287 32 6 expands expand VERB ajst-6287 32 7 samples sample NOUN ajst-6287 32 8 through through ADP ajst-6287 32 9 data data NOUN ajst-6287 32 10 generation generation NOUN ajst-6287 32 11 technologies technology NOUN ajst-6287 32 12 such such ADJ ajst-6287 32 13 as as ADP ajst-6287 32 14 picture picture NOUN ajst-6287 32 15 flipping flipping NOUN ajst-6287 32 16 , , PUNCT ajst-6287 32 17 cutting cutting NOUN ajst-6287 32 18 , , PUNCT ajst-6287 32 19 scale scale NOUN ajst-6287 32 20 change change NOUN ajst-6287 32 21 , , PUNCT ajst-6287 32 22 color color NOUN ajst-6287 32 23 jitter jitter NOUN ajst-6287 32 24 and and CCONJ ajst-6287 32 25 so so ADV ajst-6287 32 26 on on ADV ajst-6287 32 27 . . PUNCT ajst-6287 33 1 transfer transfer NOUN ajst-6287 33 2 learning learning NOUN ajst-6287 33 3 refers refer VERB ajst-6287 33 4 to to ADP ajst-6287 33 5 the the DET ajst-6287 33 6 transfer transfer NOUN ajst-6287 33 7 of of ADP ajst-6287 33 8 knowledge knowledge NOUN ajst-6287 33 9 from from ADP ajst-6287 33 10 source source NOUN ajst-6287 33 11 domain domain NOUN ajst-6287 33 12 to to PART ajst-6287 33 13 target target VERB ajst-6287 33 14 domain domain NOUN ajst-6287 33 15 . . PUNCT ajst-6287 34 1 the the DET ajst-6287 34 2 source source NOUN ajst-6287 34 3 domain domain NOUN ajst-6287 34 4 usually usually ADV ajst-6287 34 5 contains contain VERB ajst-6287 34 6 large large ADJ ajst-6287 34 7 amounts amount NOUN ajst-6287 34 8 of of ADP ajst-6287 34 9 data datum NOUN ajst-6287 34 10 samples sample NOUN ajst-6287 34 11 related relate VERB ajst-6287 34 12 or or CCONJ ajst-6287 34 13 similar similar ADJ ajst-6287 34 14 to to ADP ajst-6287 34 15 the the DET ajst-6287 34 16 target target NOUN ajst-6287 34 17 domain domain NOUN ajst-6287 34 18 . . PUNCT ajst-6287 35 1 the the DET ajst-6287 35 2 underlying underlie VERB ajst-6287 35 3 features feature NOUN ajst-6287 35 4 of of ADP ajst-6287 35 5 the the DET ajst-6287 35 6 target target NOUN ajst-6287 35 7 domain domain NOUN ajst-6287 35 8 can can AUX ajst-6287 35 9 be be AUX ajst-6287 35 10 acquired acquire VERB ajst-6287 35 11 by by ADP ajst-6287 35 12 training training NOUN ajst-6287 35 13 on on ADP ajst-6287 35 14 the the DET ajst-6287 35 15 source source NOUN ajst-6287 35 16 domain domain NOUN ajst-6287 35 17 data datum NOUN ajst-6287 35 18 . . PUNCT ajst-6287 36 1 for for ADP ajst-6287 36 2 example example NOUN ajst-6287 36 3 , , PUNCT ajst-6287 36 4 chen chen PROPN ajst-6287 36 5 et et PROPN ajst-6287 36 6 al al PROPN ajst-6287 36 7 . . PUNCT ajst-6287 37 1 [ [ X ajst-6287 37 2 3 3 X ajst-6287 37 3 ] ] PUNCT ajst-6287 37 4 used use VERB ajst-6287 37 5 deep deep ADJ ajst-6287 37 6 transfer transfer NOUN ajst-6287 37 7 learning learning NOUN ajst-6287 37 8 method method NOUN ajst-6287 37 9 for for ADP ajst-6287 37 10 plant plant NOUN ajst-6287 37 11 leaf leaf NOUN ajst-6287 37 12 disease disease NOUN ajst-6287 37 13 classification classification NOUN ajst-6287 37 14 and and CCONJ ajst-6287 37 15 identification identification NOUN ajst-6287 37 16 . . PUNCT ajst-6287 38 1 this this DET ajst-6287 38 2 method method NOUN ajst-6287 38 3 designed design VERB ajst-6287 38 4 an an DET ajst-6287 38 5 improved improved ADJ ajst-6287 38 6 vgg vgg NOUN ajst-6287 38 7 network network PROPN ajst-6287 38 8 structure structure PROPN ajst-6287 38 9 inc inc PROPN ajst-6287 38 10 - - PROPN ajst-6287 38 11 vggn vggn PROPN ajst-6287 38 12 , , PUNCT ajst-6287 38 13 and and CCONJ ajst-6287 38 14 used use VERB ajst-6287 38 15 the the DET ajst-6287 38 16 large large ADJ ajst-6287 38 17 publicly publicly ADV ajst-6287 38 18 available available ADJ ajst-6287 38 19 database database NOUN ajst-6287 38 20 imagenet imagenet NOUN ajst-6287 38 21 for for ADP ajst-6287 38 22 pre pre NOUN ajst-6287 38 23 - - NOUN ajst-6287 38 24 training training NOUN ajst-6287 38 25 of of ADP ajst-6287 38 26 the the DET ajst-6287 38 27 network network NOUN ajst-6287 38 28 model model NOUN ajst-6287 38 29 , , PUNCT ajst-6287 38 30 which which PRON ajst-6287 38 31 finally finally ADV ajst-6287 38 32 achieved achieve VERB ajst-6287 38 33 the the DET ajst-6287 38 34 identification identification NOUN ajst-6287 38 35 accuracy accuracy NOUN ajst-6287 38 36 of of ADP ajst-6287 38 37 91.83 91.83 NUM ajst-6287 38 38 % % NOUN ajst-6287 38 39 . . PUNCT ajst-6287 39 1 metalearning metalearning NOUN ajst-6287 39 2 is be AUX ajst-6287 39 3 a a DET ajst-6287 39 4 way way NOUN ajst-6287 39 5 of of ADP ajst-6287 39 6 learning learn VERB ajst-6287 39 7 how how SCONJ ajst-6287 39 8 to to PART ajst-6287 39 9 learn learn VERB ajst-6287 39 10 , , PUNCT ajst-6287 39 11 rather rather ADV ajst-6287 39 12 than than ADP ajst-6287 39 13 directly directly ADV ajst-6287 39 14 learning learn VERB ajst-6287 39 15 the the DET ajst-6287 39 16 knowledge knowledge NOUN ajst-6287 39 17 itself itself PRON ajst-6287 39 18 . . PUNCT ajst-6287 40 1 when when SCONJ ajst-6287 40 2 the the DET ajst-6287 40 3 model model NOUN ajst-6287 40 4 has have VERB ajst-6287 40 5 the the DET ajst-6287 40 6 ability ability NOUN ajst-6287 40 7 of of ADP ajst-6287 40 8 learning learn VERB ajst-6287 40 9 , , PUNCT ajst-6287 40 10 it it PRON ajst-6287 40 11 can can AUX ajst-6287 40 12 quickly quickly ADV ajst-6287 40 13 master master VERB ajst-6287 40 14 the the DET ajst-6287 40 15 ability ability NOUN ajst-6287 40 16 of of ADP ajst-6287 40 17 classification classification NOUN ajst-6287 40 18 by by ADP ajst-6287 40 19 only only ADV ajst-6287 40 20 a a DET ajst-6287 40 21 small small ADJ ajst-6287 40 22 number number NOUN ajst-6287 40 23 of of ADP ajst-6287 40 24 training training NOUN ajst-6287 40 25 samples sample NOUN ajst-6287 40 26 . . PUNCT ajst-6287 41 1 thus thus ADV ajst-6287 41 2 metalearning metalearne VERB ajst-6287 41 3 is be AUX ajst-6287 41 4 a a DET ajst-6287 41 5 learning learning NOUN ajst-6287 41 6 way way ADV ajst-6287 41 7 closer close ADV ajst-6287 41 8 to to ADP ajst-6287 41 9 humans human NOUN ajst-6287 41 10 . . PUNCT ajst-6287 42 1 in in ADP ajst-6287 42 2 this this DET ajst-6287 42 3 study study NOUN ajst-6287 42 4 we we PRON ajst-6287 42 5 propose propose VERB ajst-6287 42 6 a a DET ajst-6287 42 7 plant plant NOUN ajst-6287 42 8 leaf leaf NOUN ajst-6287 42 9 disease disease NOUN ajst-6287 42 10 classification classification NOUN ajst-6287 42 11 method method NOUN ajst-6287 42 12 based base VERB ajst-6287 42 13 on on ADP ajst-6287 42 14 fsl fsl PROPN ajst-6287 42 15 , , PUNCT ajst-6287 42 16 which which PRON ajst-6287 42 17 is be AUX ajst-6287 42 18 realized realize VERB ajst-6287 42 19 by by ADP ajst-6287 42 20 the the DET ajst-6287 42 21 way way NOUN ajst-6287 42 22 of of ADP ajst-6287 42 23 meta meta NOUN ajst-6287 42 24 - - PUNCT ajst-6287 42 25 learning learning NOUN ajst-6287 42 26 . . PUNCT ajst-6287 43 1 the the DET ajst-6287 43 2 effectiveness effectiveness NOUN ajst-6287 43 3 of of ADP ajst-6287 43 4 this this DET ajst-6287 43 5 method method NOUN ajst-6287 43 6 in in ADP ajst-6287 43 7 the the DET ajst-6287 43 8 application application NOUN ajst-6287 43 9 of of ADP ajst-6287 43 10 plant plant NOUN ajst-6287 43 11 leaf leaf NOUN ajst-6287 43 12 disease disease NOUN ajst-6287 43 13 classification classification NOUN ajst-6287 43 14 are be AUX ajst-6287 43 15 verified verify VERB ajst-6287 43 16 through through ADP ajst-6287 43 17 experiments experiment NOUN ajst-6287 43 18 . . PUNCT ajst-6287 44 1 2 2 X ajst-6287 44 2 . . X ajst-6287 44 3 research research NOUN ajst-6287 44 4 method method NOUN ajst-6287 44 5 this this DET ajst-6287 44 6 study study NOUN ajst-6287 44 7 proposes propose VERB ajst-6287 44 8 a a DET ajst-6287 44 9 classification classification NOUN ajst-6287 44 10 method method NOUN ajst-6287 44 11 of of ADP ajst-6287 44 12 plant plant NOUN ajst-6287 44 13 leaf leaf NOUN ajst-6287 44 14 diseases disease NOUN ajst-6287 44 15 based base VERB ajst-6287 44 16 on on ADP ajst-6287 44 17 fsl fsl PROPN ajst-6287 44 18 , , PUNCT ajst-6287 44 19 which which PRON ajst-6287 44 20 is be AUX ajst-6287 44 21 introduced introduce VERB ajst-6287 44 22 from from ADP ajst-6287 44 23 the the DET ajst-6287 44 24 aspects aspect NOUN ajst-6287 44 25 of of ADP ajst-6287 44 26 dataset dataset NOUN ajst-6287 44 27 , , PUNCT ajst-6287 44 28 feature feature NOUN ajst-6287 44 29 extraction extraction NOUN ajst-6287 44 30 and and CCONJ ajst-6287 44 31 meta meta NOUN ajst-6287 44 32 - - PUNCT ajst-6287 44 33 learning learning NOUN ajst-6287 44 34 . . PUNCT ajst-6287 45 1 2.1 2.1 NUM ajst-6287 45 2 . . PUNCT ajst-6287 45 3 dataset dataset VERB ajst-6287 45 4 traditional traditional ADJ ajst-6287 45 5 deep deep ADJ ajst-6287 45 6 learning learning NOUN ajst-6287 45 7 divides divide VERB ajst-6287 45 8 the the DET ajst-6287 45 9 required require VERB ajst-6287 45 10 data datum NOUN ajst-6287 45 11 into into ADP ajst-6287 45 12 training training NOUN ajst-6287 45 13 set set NOUN ajst-6287 45 14 , , PUNCT ajst-6287 45 15 validation validation NOUN ajst-6287 45 16 set set NOUN ajst-6287 45 17 , , PUNCT ajst-6287 45 18 and and CCONJ ajst-6287 45 19 test test NOUN ajst-6287 45 20 set set VERB ajst-6287 45 21 . . PUNCT ajst-6287 46 1 differently differently ADV ajst-6287 46 2 metalearning metalearne VERB ajst-6287 46 3 is be AUX ajst-6287 46 4 driven drive VERB ajst-6287 46 5 by by ADP ajst-6287 46 6 task task NOUN ajst-6287 46 7 rather rather ADV ajst-6287 46 8 than than ADP ajst-6287 46 9 data datum NOUN ajst-6287 46 10 . . PUNCT ajst-6287 47 1 each each DET ajst-6287 47 2 task task NOUN ajst-6287 47 3 contains contain VERB ajst-6287 47 4 the the DET ajst-6287 47 5 support support NOUN ajst-6287 47 6 set set VERB ajst-6287 47 7 and and CCONJ ajst-6287 47 8 the the DET ajst-6287 47 9 query query NOUN ajst-6287 47 10 set set VERB ajst-6287 47 11 , , PUNCT ajst-6287 47 12 with with ADP ajst-6287 47 13 the the DET ajst-6287 47 14 former former ADJ ajst-6287 47 15 providing provide VERB ajst-6287 47 16 a a DET ajst-6287 47 17 small small ADJ ajst-6287 47 18 amount amount NOUN ajst-6287 47 19 of of ADP ajst-6287 47 20 data datum NOUN ajst-6287 47 21 for for ADP ajst-6287 47 22 model model NOUN ajst-6287 47 23 tuning tuning NOUN ajst-6287 47 24 , , PUNCT ajst-6287 47 25 and and CCONJ ajst-6287 47 26 the the DET ajst-6287 47 27 latter latter ADJ ajst-6287 47 28 being be AUX ajst-6287 47 29 83 83 NUM ajst-6287 47 30 used use VERB ajst-6287 47 31 to to PART ajst-6287 47 32 verify verify VERB ajst-6287 47 33 the the DET ajst-6287 47 34 generalization generalization NOUN ajst-6287 47 35 ability ability NOUN ajst-6287 47 36 of of ADP ajst-6287 47 37 the the DET ajst-6287 47 38 model model NOUN ajst-6287 47 39 on on ADP ajst-6287 47 40 that that DET ajst-6287 47 41 task task NOUN ajst-6287 47 42 . . PUNCT ajst-6287 48 1 2.2 2.2 NUM ajst-6287 48 2 . . PUNCT ajst-6287 49 1 feature feature NOUN ajst-6287 49 2 extraction extraction NOUN ajst-6287 49 3 in in ADP ajst-6287 49 4 this this DET ajst-6287 49 5 study study NOUN ajst-6287 49 6 , , PUNCT ajst-6287 49 7 the the DET ajst-6287 49 8 network network NOUN ajst-6287 49 9 structure structure NOUN ajst-6287 49 10 before before ADP ajst-6287 49 11 the the DET ajst-6287 49 12 fully fully ADV ajst-6287 49 13 connected connected ADJ ajst-6287 49 14 layer layer NOUN ajst-6287 49 15 in in ADP ajst-6287 49 16 the the DET ajst-6287 49 17 inception inception ADJ ajst-6287 49 18 v3 v3 PROPN ajst-6287 49 19 model model NOUN ajst-6287 50 1 [ [ X ajst-6287 50 2 4 4 NUM ajst-6287 50 3 ] ] PUNCT ajst-6287 50 4 was be AUX ajst-6287 50 5 used use VERB ajst-6287 50 6 for for ADP ajst-6287 50 7 feature feature NOUN ajst-6287 50 8 extraction extraction NOUN ajst-6287 50 9 . . PUNCT ajst-6287 51 1 the the DET ajst-6287 51 2 model model NOUN ajst-6287 51 3 decomposes decompose VERB ajst-6287 51 4 the the DET ajst-6287 51 5 larger large ADJ ajst-6287 51 6 convolutional convolutional ADJ ajst-6287 51 7 kernel kernel NOUN ajst-6287 51 8 in in ADP ajst-6287 51 9 the the DET ajst-6287 51 10 traditional traditional ADJ ajst-6287 51 11 network network NOUN ajst-6287 51 12 hierarchy hierarchy NOUN ajst-6287 51 13 to to ADP ajst-6287 51 14 reducing reduce VERB ajst-6287 51 15 the the DET ajst-6287 51 16 number number NOUN ajst-6287 51 17 of of ADP ajst-6287 51 18 parameters parameter NOUN ajst-6287 51 19 of of ADP ajst-6287 51 20 the the DET ajst-6287 51 21 model model NOUN ajst-6287 51 22 . . PUNCT ajst-6287 52 1 the the DET ajst-6287 52 2 inception inception NOUN ajst-6287 52 3 module module NOUN ajst-6287 52 4 group group NOUN ajst-6287 52 5 is be AUX ajst-6287 52 6 used use VERB ajst-6287 52 7 to to PART ajst-6287 52 8 balance balance VERB ajst-6287 52 9 the the DET ajst-6287 52 10 depth depth NOUN ajst-6287 52 11 and and CCONJ ajst-6287 52 12 width width NOUN ajst-6287 52 13 of of ADP ajst-6287 52 14 the the DET ajst-6287 52 15 network network NOUN ajst-6287 52 16 hierarchy hierarchy NOUN ajst-6287 52 17 , , PUNCT ajst-6287 52 18 which which PRON ajst-6287 52 19 improves improve VERB ajst-6287 52 20 the the DET ajst-6287 52 21 ability ability NOUN ajst-6287 52 22 of of ADP ajst-6287 52 23 feature feature NOUN ajst-6287 52 24 extraction extraction NOUN ajst-6287 52 25 . . PUNCT ajst-6287 53 1 in in ADP ajst-6287 53 2 addition addition NOUN ajst-6287 53 3 , , PUNCT ajst-6287 53 4 it it PRON ajst-6287 53 5 uses use VERB ajst-6287 53 6 batch batch NOUN ajst-6287 53 7 normalization normalization NOUN ajst-6287 53 8 to to PART ajst-6287 53 9 alleviate alleviate VERB ajst-6287 53 10 the the DET ajst-6287 53 11 problem problem NOUN ajst-6287 53 12 of of ADP ajst-6287 53 13 gradient gradient ADJ ajst-6287 53 14 disappearance disappearance NOUN ajst-6287 53 15 in in ADP ajst-6287 53 16 the the DET ajst-6287 53 17 deep deep ADJ ajst-6287 53 18 network network NOUN ajst-6287 53 19 hierarchy hierarchy NOUN ajst-6287 53 20 . . PUNCT ajst-6287 54 1 2.3 2.3 NUM ajst-6287 54 2 . . PUNCT ajst-6287 55 1 meta meta VERB ajst-6287 55 2 - - PUNCT ajst-6287 55 3 learning learn VERB ajst-6287 55 4 the the DET ajst-6287 55 5 meta meta ADV ajst-6287 55 6 - - PUNCT ajst-6287 55 7 learning learn VERB ajst-6287 55 8 method method NOUN ajst-6287 55 9 mainly mainly ADV ajst-6287 55 10 includes include VERB ajst-6287 55 11 two two NUM ajst-6287 55 12 steps step NOUN ajst-6287 55 13 : : PUNCT ajst-6287 55 14 meta meta ADJ ajst-6287 55 15 - - PUNCT ajst-6287 55 16 training training NOUN ajst-6287 55 17 and and CCONJ ajst-6287 55 18 meta meta NOUN ajst-6287 55 19 - - PUNCT ajst-6287 55 20 testing testing NOUN ajst-6287 55 21 : : PUNCT ajst-6287 55 22 ( ( PUNCT ajst-6287 55 23 1 1 X ajst-6287 55 24 ) ) PUNCT ajst-6287 55 25 meta meta NOUN ajst-6287 55 26 - - PUNCT ajst-6287 55 27 training training NOUN ajst-6287 55 28 in in ADP ajst-6287 55 29 this this DET ajst-6287 55 30 stage stage NOUN ajst-6287 55 31 , , PUNCT ajst-6287 55 32 the the DET ajst-6287 55 33 meta meta ADJ ajst-6287 55 34 - - PUNCT ajst-6287 55 35 training training NOUN ajst-6287 55 36 data datum NOUN ajst-6287 55 37 set set VERB ajst-6287 55 38 is be AUX ajst-6287 55 39 organized organize VERB ajst-6287 55 40 into into ADP ajst-6287 55 41 tasks task NOUN ajst-6287 55 42 for for ADP ajst-6287 55 43 model model NOUN ajst-6287 55 44 training training NOUN ajst-6287 55 45 . . PUNCT ajst-6287 56 1 n n PRON ajst-6287 56 2 categories category NOUN ajst-6287 56 3 were be AUX ajst-6287 56 4 randomly randomly ADV ajst-6287 56 5 selected select VERB ajst-6287 56 6 for for ADP ajst-6287 56 7 each each DET ajst-6287 56 8 task task NOUN ajst-6287 56 9 , , PUNCT ajst-6287 56 10 and and CCONJ ajst-6287 56 11 t t PROPN ajst-6287 56 12 data datum NOUN ajst-6287 56 13 were be AUX ajst-6287 56 14 randomly randomly ADV ajst-6287 56 15 sampled sample VERB ajst-6287 56 16 in in ADP ajst-6287 56 17 each each DET ajst-6287 56 18 category category NOUN ajst-6287 56 19 . . PUNCT ajst-6287 57 1 the the DET ajst-6287 57 2 former former ADJ ajst-6287 57 3 k k PROPN ajst-6287 57 4 data datum NOUN ajst-6287 57 5 constitute constitute VERB ajst-6287 57 6 the the DET ajst-6287 57 7 support support NOUN ajst-6287 57 8 set set NOUN ajst-6287 57 9 , , PUNCT ajst-6287 57 10 expressed express VERB ajst-6287 57 11 as as ADP ajst-6287 57 12 s s NOUN ajst-6287 57 13 π‘₯ π‘₯ X ajst-6287 57 14 , , PUNCT ajst-6287 57 15 𝑦 𝑦 NOUN ajst-6287 57 16 βˆ— βˆ— NOUN ajst-6287 57 17 , , PUNCT ajst-6287 57 18 which which PRON ajst-6287 57 19 is be AUX ajst-6287 57 20 often often ADV ajst-6287 57 21 called call VERB ajst-6287 57 22 n n CCONJ ajst-6287 57 23 - - PUNCT ajst-6287 57 24 way way NOUN ajst-6287 57 25 k k NOUN ajst-6287 57 26 - - NOUN ajst-6287 57 27 shot shot NOUN ajst-6287 57 28 in in ADP ajst-6287 57 29 fsl fsl PROPN ajst-6287 57 30 . . PUNCT ajst-6287 58 1 the the DET ajst-6287 58 2 remaining remain VERB ajst-6287 58 3 ( ( PUNCT ajst-6287 58 4 t t PROPN ajst-6287 58 5 - - PUNCT ajst-6287 58 6 k k NOUN ajst-6287 58 7 ) ) PUNCT ajst-6287 58 8 data datum NOUN ajst-6287 58 9 of of ADP ajst-6287 58 10 each each DET ajst-6287 58 11 category category NOUN ajst-6287 58 12 constitute constitute VERB ajst-6287 58 13 the the DET ajst-6287 58 14 query query NOUN ajst-6287 58 15 set set NOUN ajst-6287 58 16 , , PUNCT ajst-6287 58 17 expressed express VERB ajst-6287 58 18 as as ADP ajst-6287 58 19 q q PROPN ajst-6287 58 20 π‘₯ π‘₯ PROPN ajst-6287 58 21 , , PUNCT ajst-6287 58 22 𝑦 𝑦 NOUN ajst-6287 58 23 βˆ— βˆ— NOUN ajst-6287 58 24 . . PUNCT ajst-6287 59 1 the the DET ajst-6287 59 2 support support NOUN ajst-6287 59 3 set set VERB ajst-6287 59 4 in in ADP ajst-6287 59 5 each each DET ajst-6287 59 6 task task NOUN ajst-6287 59 7 is be AUX ajst-6287 59 8 used use VERB ajst-6287 59 9 to to PART ajst-6287 59 10 enable enable VERB ajst-6287 59 11 the the DET ajst-6287 59 12 model model NOUN ajst-6287 59 13 to to PART ajst-6287 59 14 establish establish VERB ajst-6287 59 15 a a DET ajst-6287 59 16 classification classification NOUN ajst-6287 59 17 pattern pattern NOUN ajst-6287 59 18 , , PUNCT ajst-6287 59 19 while while SCONJ ajst-6287 59 20 the the DET ajst-6287 59 21 query query NOUN ajst-6287 59 22 set set NOUN ajst-6287 59 23 is be AUX ajst-6287 59 24 used use VERB ajst-6287 59 25 to to PART ajst-6287 59 26 verify verify VERB ajst-6287 59 27 the the DET ajst-6287 59 28 effect effect NOUN ajst-6287 59 29 of of ADP ajst-6287 59 30 this this DET ajst-6287 59 31 pattern pattern NOUN ajst-6287 59 32 . . PUNCT ajst-6287 60 1 in in ADP ajst-6287 60 2 this this DET ajst-6287 60 3 study study NOUN ajst-6287 60 4 , , PUNCT ajst-6287 60 5 the the DET ajst-6287 60 6 prototype prototype NOUN ajst-6287 60 7 method method NOUN ajst-6287 60 8 [ [ X ajst-6287 60 9 5 5 NUM ajst-6287 60 10 ] ] PUNCT ajst-6287 60 11 was be AUX ajst-6287 60 12 used use VERB ajst-6287 60 13 to to PART ajst-6287 60 14 support support VERB ajst-6287 60 15 the the DET ajst-6287 60 16 establishment establishment NOUN ajst-6287 60 17 of of ADP ajst-6287 60 18 the the DET ajst-6287 60 19 classification classification NOUN ajst-6287 60 20 patterns pattern NOUN ajst-6287 60 21 . . PUNCT ajst-6287 61 1 for for ADP ajst-6287 61 2 an an DET ajst-6287 61 3 n n NUM ajst-6287 61 4 - - PUNCT ajst-6287 61 5 way way NOUN ajst-6287 61 6 k k NOUN ajst-6287 61 7 - - PUNCT ajst-6287 61 8 shot shoot VERB ajst-6287 61 9 task task NOUN ajst-6287 61 10 , , PUNCT ajst-6287 61 11 the the DET ajst-6287 61 12 prototype prototype NOUN ajst-6287 61 13 feature feature NOUN ajst-6287 61 14 𝑐 𝑐 PROPN ajst-6287 61 15 of of ADP ajst-6287 61 16 the the DET ajst-6287 61 17 category category NOUN ajst-6287 61 18 k k PROPN ajst-6287 61 19 can can AUX ajst-6287 61 20 be be AUX ajst-6287 61 21 expressed express VERB ajst-6287 61 22 as as ADP ajst-6287 61 23 : : PUNCT ajst-6287 61 24 c c NOUN ajst-6287 61 25 1 1 NUM ajst-6287 61 26 𝐾 𝐾 NOUN ajst-6287 61 27 π‘“βˆ… π‘“βˆ… PUNCT ajst-6287 61 28 π‘₯ π‘₯ X ajst-6287 61 29 where where SCONJ ajst-6287 61 30 x x PRON ajst-6287 61 31 represents represent VERB ajst-6287 61 32 the the DET ajst-6287 61 33 sample sample NOUN ajst-6287 61 34 data datum NOUN ajst-6287 61 35 of of ADP ajst-6287 61 36 category category NOUN ajst-6287 61 37 k k NOUN ajst-6287 61 38 , , PUNCT ajst-6287 61 39 π‘“βˆ… π‘“βˆ… PUNCT ajst-6287 61 40 represents represent VERB ajst-6287 61 41 a a DET ajst-6287 61 42 functional functional ADJ ajst-6287 61 43 map map NOUN ajst-6287 61 44 with with ADP ajst-6287 61 45 parameter parameter NOUN ajst-6287 61 46 βˆ… βˆ… NOUN ajst-6287 61 47 , , PUNCT ajst-6287 61 48 which which PRON ajst-6287 61 49 is be AUX ajst-6287 61 50 automatically automatically ADV ajst-6287 61 51 learned learn VERB ajst-6287 61 52 by by ADP ajst-6287 61 53 the the DET ajst-6287 61 54 neural neural ADJ ajst-6287 61 55 network network NOUN ajst-6287 61 56 model model NOUN ajst-6287 61 57 . . PUNCT ajst-6287 62 1 samples sample NOUN ajst-6287 62 2 in in ADP ajst-6287 62 3 the the DET ajst-6287 62 4 query query NOUN ajst-6287 62 5 set set VERB ajst-6287 62 6 use use VERB ajst-6287 62 7 the the DET ajst-6287 62 8 same same ADJ ajst-6287 62 9 functional functional ADJ ajst-6287 62 10 map map NOUN ajst-6287 62 11 for for ADP ajst-6287 62 12 feature feature NOUN ajst-6287 62 13 extraction extraction NOUN ajst-6287 62 14 and and CCONJ ajst-6287 62 15 measure measure VERB ajst-6287 62 16 the the DET ajst-6287 62 17 distance distance NOUN ajst-6287 62 18 from from ADP ajst-6287 62 19 each each DET ajst-6287 62 20 prototype prototype NOUN ajst-6287 62 21 . . PUNCT ajst-6287 63 1 in in ADP ajst-6287 63 2 this this DET ajst-6287 63 3 study study NOUN ajst-6287 63 4 the the DET ajst-6287 63 5 distance distance NOUN ajst-6287 63 6 was be AUX ajst-6287 63 7 measured measure VERB ajst-6287 63 8 by by ADP ajst-6287 63 9 euclidean euclidean ADJ ajst-6287 63 10 distance distance NOUN ajst-6287 63 11 . . PUNCT ajst-6287 64 1 given give VERB ajst-6287 64 2 two two NUM ajst-6287 64 3 n n CCONJ ajst-6287 64 4 - - PUNCT ajst-6287 64 5 dimensional dimensional ADJ ajst-6287 64 6 vectors vector NOUN ajst-6287 64 7 x x X ajst-6287 64 8 π‘₯ π‘₯ NOUN ajst-6287 64 9 π‘₯ π‘₯ X ajst-6287 64 10 … … PUNCT ajst-6287 64 11 π‘₯ π‘₯ PROPN ajst-6287 64 12 and and CCONJ ajst-6287 64 13 y y PROPN ajst-6287 64 14 𝑦 𝑦 PROPN ajst-6287 64 15 𝑦 𝑦 NOUN ajst-6287 64 16 … … PUNCT ajst-6287 64 17 𝑦 𝑦 NOUN ajst-6287 64 18 , , PUNCT ajst-6287 64 19 their their PRON ajst-6287 64 20 euclidean euclidean ADJ ajst-6287 64 21 distance distance NOUN ajst-6287 64 22 can can AUX ajst-6287 64 23 be be AUX ajst-6287 64 24 expressed express VERB ajst-6287 64 25 as as ADP ajst-6287 64 26 : : PUNCT ajst-6287 64 27 d d NOUN ajst-6287 64 28 x x PROPN ajst-6287 64 29 , , PUNCT ajst-6287 64 30 y y PROPN ajst-6287 64 31 π‘₯ π‘₯ X ajst-6287 64 32 𝑦 𝑦 NOUN ajst-6287 64 33 for for ADP ajst-6287 64 34 a a DET ajst-6287 64 35 sample sample NOUN ajst-6287 64 36 x x PUNCT ajst-6287 64 37 with with ADP ajst-6287 64 38 a a DET ajst-6287 64 39 label label NOUN ajst-6287 64 40 k k NOUN ajst-6287 64 41 in in ADP ajst-6287 64 42 the the DET ajst-6287 64 43 query query NOUN ajst-6287 64 44 set set NOUN ajst-6287 64 45 , , PUNCT ajst-6287 64 46 the the DET ajst-6287 64 47 euclidean euclidean ADJ ajst-6287 64 48 distance distance NOUN ajst-6287 64 49 from from ADP ajst-6287 64 50 each each DET ajst-6287 64 51 prototype prototype NOUN ajst-6287 64 52 feature feature NOUN ajst-6287 64 53 is be AUX ajst-6287 64 54 calculated calculate VERB ajst-6287 64 55 . . PUNCT ajst-6287 65 1 then then ADV ajst-6287 65 2 the the DET ajst-6287 65 3 probability probability NOUN ajst-6287 65 4 distribution distribution NOUN ajst-6287 65 5 of of ADP ajst-6287 65 6 the the DET ajst-6287 65 7 classification classification NOUN ajst-6287 65 8 can can AUX ajst-6287 65 9 be be AUX ajst-6287 65 10 further far ADV ajst-6287 65 11 calculated calculate VERB ajst-6287 65 12 through through ADP ajst-6287 65 13 the the DET ajst-6287 65 14 softmax softmax NOUN ajst-6287 65 15 function function NOUN ajst-6287 65 16 . . PUNCT ajst-6287 66 1 the the DET ajst-6287 66 2 probability probability NOUN ajst-6287 66 3 of of ADP ajst-6287 66 4 x x PRON ajst-6287 66 5 being be AUX ajst-6287 66 6 classified classify VERB ajst-6287 66 7 into into ADP ajst-6287 66 8 category category NOUN ajst-6287 66 9 k k PROPN ajst-6287 66 10 can can AUX ajst-6287 66 11 be be AUX ajst-6287 66 12 expressed express VERB ajst-6287 66 13 as as ADP ajst-6287 66 14 : : PUNCT ajst-6287 66 15 π‘ƒβˆ… π‘ƒβˆ… PROPN ajst-6287 66 16 y y PROPN ajst-6287 66 17 k|x k|x PROPN ajst-6287 66 18 exp exp PROPN ajst-6287 66 19 𝑑 𝑑 PROPN ajst-6287 66 20 π‘“βˆ… π‘“βˆ… ADP ajst-6287 66 21 π‘₯ π‘₯ PROPN ajst-6287 66 22 , , PUNCT ajst-6287 66 23 𝑐 𝑐 PROPN ajst-6287 66 24 βˆ‘ βˆ‘ PUNCT ajst-6287 66 25 exp exp PROPN ajst-6287 66 26 𝑑 𝑑 PROPN ajst-6287 66 27 π‘“βˆ… π‘“βˆ… ADP ajst-6287 66 28 π‘₯ π‘₯ PROPN ajst-6287 66 29 , , PUNCT ajst-6287 66 30 𝑐 𝑐 VERB ajst-6287 66 31 the the DET ajst-6287 66 32 purpose purpose NOUN ajst-6287 66 33 of of ADP ajst-6287 66 34 the the DET ajst-6287 66 35 meta meta ADJ ajst-6287 66 36 - - PUNCT ajst-6287 66 37 training training NOUN ajst-6287 66 38 stage stage NOUN ajst-6287 66 39 is be AUX ajst-6287 66 40 to to PART ajst-6287 66 41 enable enable VERB ajst-6287 66 42 the the DET ajst-6287 66 43 prototype prototype NOUN ajst-6287 66 44 features feature VERB ajst-6287 66 45 to to PART ajst-6287 66 46 represent represent VERB ajst-6287 66 47 the the DET ajst-6287 66 48 corresponding correspond VERB ajst-6287 66 49 category category NOUN ajst-6287 66 50 as as ADV ajst-6287 66 51 accurately accurately ADV ajst-6287 66 52 as as ADP ajst-6287 66 53 possible possible ADJ ajst-6287 66 54 , , PUNCT ajst-6287 66 55 and and CCONJ ajst-6287 66 56 to to PART ajst-6287 66 57 maximize maximize VERB ajst-6287 66 58 the the DET ajst-6287 66 59 probability probability NOUN ajst-6287 66 60 of of ADP ajst-6287 66 61 predicting predict VERB ajst-6287 66 62 the the DET ajst-6287 66 63 classification classification NOUN ajst-6287 66 64 result result NOUN ajst-6287 66 65 of of ADP ajst-6287 66 66 k. k. PROPN ajst-6287 66 67 therefore therefore ADV ajst-6287 66 68 , , PUNCT ajst-6287 66 69 the the DET ajst-6287 66 70 final final ADJ ajst-6287 66 71 loss loss NOUN ajst-6287 66 72 function function NOUN ajst-6287 66 73 is be AUX ajst-6287 66 74 defined define VERB ajst-6287 66 75 as as SCONJ ajst-6287 66 76 follows follow VERB ajst-6287 66 77 : : PUNCT ajst-6287 66 78 j j PROPN ajst-6287 66 79 βˆ… βˆ… NOUN ajst-6287 66 80 logπ‘ƒβˆ… logπ‘ƒβˆ… NOUN ajst-6287 66 81 𝑦 𝑦 NOUN ajst-6287 66 82 π‘˜|π‘₯ π‘˜|π‘₯ X ajst-6287 66 83 ( ( PUNCT ajst-6287 66 84 2 2 X ajst-6287 66 85 ) ) PUNCT ajst-6287 66 86 meta meta NOUN ajst-6287 66 87 - - PUNCT ajst-6287 66 88 testing testing NOUN ajst-6287 66 89 after after ADP ajst-6287 66 90 the the DET ajst-6287 66 91 meta meta ADJ ajst-6287 66 92 - - PUNCT ajst-6287 66 93 training training NOUN ajst-6287 66 94 process process NOUN ajst-6287 66 95 , , PUNCT ajst-6287 66 96 the the DET ajst-6287 66 97 network network NOUN ajst-6287 66 98 model model NOUN ajst-6287 66 99 has have VERB ajst-6287 66 100 the the DET ajst-6287 66 101 ability ability NOUN ajst-6287 66 102 of of ADP ajst-6287 66 103 few few ADJ ajst-6287 66 104 - - PUNCT ajst-6287 66 105 shot shot NOUN ajst-6287 66 106 learning learning NOUN ajst-6287 66 107 . . PUNCT ajst-6287 67 1 in in ADP ajst-6287 67 2 the the DET ajst-6287 67 3 meta meta ADJ ajst-6287 67 4 - - PUNCT ajst-6287 67 5 test test NOUN ajst-6287 67 6 dataset dataset NOUN ajst-6287 67 7 , , PUNCT ajst-6287 67 8 the the DET ajst-6287 67 9 prototype prototype NOUN ajst-6287 67 10 of of ADP ajst-6287 67 11 each each DET ajst-6287 67 12 category category NOUN ajst-6287 67 13 of of ADP ajst-6287 67 14 the the DET ajst-6287 67 15 support support NOUN ajst-6287 67 16 set set NOUN ajst-6287 67 17 is be AUX ajst-6287 67 18 calculated calculate VERB ajst-6287 67 19 through through ADP ajst-6287 67 20 the the DET ajst-6287 67 21 feature feature NOUN ajst-6287 67 22 extraction extraction NOUN ajst-6287 67 23 module module NOUN ajst-6287 67 24 . . PUNCT ajst-6287 68 1 then then ADV ajst-6287 68 2 the the DET ajst-6287 68 3 category category NOUN ajst-6287 68 4 of of ADP ajst-6287 68 5 each each DET ajst-6287 68 6 sample sample NOUN ajst-6287 68 7 in in ADP ajst-6287 68 8 the the DET ajst-6287 68 9 query query NOUN ajst-6287 68 10 set set NOUN ajst-6287 68 11 was be AUX ajst-6287 68 12 predicted predict VERB ajst-6287 68 13 to to PART ajst-6287 68 14 test test VERB ajst-6287 68 15 the the DET ajst-6287 68 16 effect effect NOUN ajst-6287 68 17 of of ADP ajst-6287 68 18 the the DET ajst-6287 68 19 trained train VERB ajst-6287 68 20 model model NOUN ajst-6287 68 21 . . PUNCT ajst-6287 69 1 3 3 X ajst-6287 69 2 . . X ajst-6287 69 3 experimental experimental ADJ ajst-6287 69 4 results result NOUN ajst-6287 69 5 and and CCONJ ajst-6287 69 6 analysis analysis NOUN ajst-6287 69 7 to to PART ajst-6287 69 8 verify verify VERB ajst-6287 69 9 the the DET ajst-6287 69 10 algorithm algorithm NOUN ajst-6287 69 11 of of ADP ajst-6287 69 12 this this DET ajst-6287 69 13 study study NOUN ajst-6287 69 14 , , PUNCT ajst-6287 69 15 several several ADJ ajst-6287 69 16 experiments experiment NOUN ajst-6287 69 17 were be AUX ajst-6287 69 18 performed perform VERB ajst-6287 69 19 . . PUNCT ajst-6287 70 1 the the DET ajst-6287 70 2 model model NOUN ajst-6287 70 3 was be AUX ajst-6287 70 4 optimized optimize VERB ajst-6287 70 5 by by ADP ajst-6287 70 6 using use VERB ajst-6287 70 7 a a DET ajst-6287 70 8 minibatch minibatch NOUN ajst-6287 70 9 stochastic stochastic ADJ ajst-6287 70 10 gradient gradient ADJ ajst-6287 70 11 descent descent NOUN ajst-6287 70 12 algorithm algorithm NOUN ajst-6287 70 13 . . PUNCT ajst-6287 71 1 the the DET ajst-6287 71 2 publicly publicly ADV ajst-6287 71 3 available available ADJ ajst-6287 71 4 dataset dataset NOUN ajst-6287 71 5 plantvillage plantvillage NOUN ajst-6287 71 6 [ [ X ajst-6287 71 7 6 6 NUM ajst-6287 71 8 ] ] PUNCT ajst-6287 71 9 was be AUX ajst-6287 71 10 used use VERB ajst-6287 71 11 for for ADP ajst-6287 71 12 plant plant NOUN ajst-6287 71 13 leaf leaf NOUN ajst-6287 71 14 disease disease NOUN ajst-6287 71 15 classification classification NOUN ajst-6287 71 16 experiments experiment NOUN ajst-6287 71 17 . . PUNCT ajst-6287 72 1 the the DET ajst-6287 72 2 dataset dataset NOUN ajst-6287 72 3 contains contain VERB ajst-6287 72 4 54303 54303 NUM ajst-6287 72 5 images image NOUN ajst-6287 72 6 of of ADP ajst-6287 72 7 plant plant NOUN ajst-6287 72 8 leaves leave NOUN ajst-6287 72 9 , , PUNCT ajst-6287 72 10 covering cover VERB ajst-6287 72 11 14 14 NUM ajst-6287 72 12 plants plant NOUN ajst-6287 72 13 and and CCONJ ajst-6287 72 14 26 26 NUM ajst-6287 72 15 leaf leaf NOUN ajst-6287 72 16 disease disease NOUN ajst-6287 72 17 types type NOUN ajst-6287 72 18 , , PUNCT ajst-6287 72 19 forming form VERB ajst-6287 72 20 38 38 NUM ajst-6287 72 21 categories category NOUN ajst-6287 72 22 based base VERB ajst-6287 72 23 on on ADP ajst-6287 72 24 the the DET ajst-6287 72 25 combination combination NOUN ajst-6287 72 26 of of ADP ajst-6287 72 27 plant plant NOUN ajst-6287 72 28 and and CCONJ ajst-6287 72 29 leaf leaf NOUN ajst-6287 72 30 disease disease NOUN ajst-6287 72 31 types type NOUN ajst-6287 72 32 ( ( PUNCT ajst-6287 72 33 including include VERB ajst-6287 72 34 healthy healthy ADJ ajst-6287 72 35 type type NOUN ajst-6287 72 36 ) ) PUNCT ajst-6287 72 37 . . PUNCT ajst-6287 73 1 in in ADP ajst-6287 73 2 this this DET ajst-6287 73 3 study study NOUN ajst-6287 73 4 , , PUNCT ajst-6287 73 5 samples sample NOUN ajst-6287 73 6 from from ADP ajst-6287 73 7 28 28 NUM ajst-6287 73 8 categories category NOUN ajst-6287 73 9 were be AUX ajst-6287 73 10 randomly randomly ADV ajst-6287 73 11 selected select VERB ajst-6287 73 12 as as ADP ajst-6287 73 13 the the DET ajst-6287 73 14 meta meta ADJ ajst-6287 73 15 - - PUNCT ajst-6287 73 16 training training NOUN ajst-6287 73 17 dataset dataset NOUN ajst-6287 73 18 , , PUNCT ajst-6287 73 19 and and CCONJ ajst-6287 73 20 the the DET ajst-6287 73 21 remaining remain VERB ajst-6287 73 22 samples sample NOUN ajst-6287 73 23 of of ADP ajst-6287 73 24 10 10 NUM ajst-6287 73 25 categories category NOUN ajst-6287 73 26 were be AUX ajst-6287 73 27 used use VERB ajst-6287 73 28 as as ADP ajst-6287 73 29 the the DET ajst-6287 73 30 meta meta ADJ ajst-6287 73 31 - - PUNCT ajst-6287 73 32 test test NOUN ajst-6287 73 33 dataset dataset NOUN ajst-6287 73 34 . . PUNCT ajst-6287 74 1 different different ADJ ajst-6287 74 2 n n CCONJ ajst-6287 74 3 - - PUNCT ajst-6287 74 4 way way NOUN ajst-6287 74 5 k k NOUN ajst-6287 74 6 - - PUNCT ajst-6287 74 7 shot shot ADJ ajst-6287 74 8 parameters parameter NOUN ajst-6287 74 9 were be AUX ajst-6287 74 10 configured configure VERB ajst-6287 74 11 to to PART ajst-6287 74 12 generate generate VERB ajst-6287 74 13 tasks task NOUN ajst-6287 74 14 during during ADP ajst-6287 74 15 the the DET ajst-6287 74 16 meta meta ADJ ajst-6287 74 17 - - PUNCT ajst-6287 74 18 training training NOUN ajst-6287 74 19 phase phase NOUN ajst-6287 74 20 in in ADP ajst-6287 74 21 the the DET ajst-6287 74 22 experiment experiment NOUN ajst-6287 74 23 . . PUNCT ajst-6287 75 1 the the DET ajst-6287 75 2 parameter parameter NOUN ajst-6287 75 3 n n CCONJ ajst-6287 75 4 in in ADP ajst-6287 75 5 the the DET ajst-6287 75 6 support support NOUN ajst-6287 75 7 set set VERB ajst-6287 75 8 takes take VERB ajst-6287 75 9 values value NOUN ajst-6287 75 10 of of ADP ajst-6287 75 11 3 3 NUM ajst-6287 75 12 and and CCONJ ajst-6287 75 13 5 5 NUM ajst-6287 75 14 , , PUNCT ajst-6287 75 15 and and CCONJ ajst-6287 75 16 the the DET ajst-6287 75 17 parameter parameter NOUN ajst-6287 75 18 k k PROPN ajst-6287 75 19 takes take VERB ajst-6287 75 20 1,5 1,5 NUM ajst-6287 75 21 and and CCONJ ajst-6287 75 22 10 10 NUM ajst-6287 75 23 . . PUNCT ajst-6287 76 1 the the DET ajst-6287 76 2 query query NOUN ajst-6287 76 3 set set NOUN ajst-6287 76 4 was be AUX ajst-6287 76 5 sampled sample VERB ajst-6287 76 6 in in ADP ajst-6287 76 7 the the DET ajst-6287 76 8 same same ADJ ajst-6287 76 9 category category NOUN ajst-6287 76 10 as as SCONJ ajst-6287 76 11 the the DET ajst-6287 76 12 support support NOUN ajst-6287 76 13 set set NOUN ajst-6287 76 14 , , PUNCT ajst-6287 76 15 with with ADP ajst-6287 76 16 15 15 NUM ajst-6287 76 17 images image NOUN ajst-6287 76 18 constantly constantly ADV ajst-6287 76 19 sampled sample VERB ajst-6287 76 20 for for ADP ajst-6287 76 21 each each DET ajst-6287 76 22 category category NOUN ajst-6287 76 23 , , PUNCT ajst-6287 76 24 but but CCONJ ajst-6287 76 25 did do AUX ajst-6287 76 26 not not PART ajst-6287 76 27 overlap overlap VERB ajst-6287 76 28 with with ADP ajst-6287 76 29 the the DET ajst-6287 76 30 support support NOUN ajst-6287 76 31 set set VERB ajst-6287 76 32 samples sample NOUN ajst-6287 76 33 . . PUNCT ajst-6287 77 1 the the DET ajst-6287 77 2 meta meta ADJ ajst-6287 77 3 - - PUNCT ajst-6287 77 4 test test NOUN ajst-6287 77 5 phase phase NOUN ajst-6287 77 6 generated generate VERB ajst-6287 77 7 500 500 NUM ajst-6287 77 8 tasks task NOUN ajst-6287 77 9 , , PUNCT ajst-6287 77 10 each each PRON ajst-6287 77 11 with with ADP ajst-6287 77 12 the the DET ajst-6287 77 13 same same ADJ ajst-6287 77 14 as as ADP ajst-6287 77 15 the the DET ajst-6287 77 16 n n NUM ajst-6287 77 17 - - PUNCT ajst-6287 77 18 way way NOUN ajst-6287 77 19 k k NOUN ajst-6287 77 20 - - PUNCT ajst-6287 77 21 shot shoot VERB ajst-6287 77 22 parameters parameter NOUN ajst-6287 77 23 in in ADP ajst-6287 77 24 the the DET ajst-6287 77 25 meta meta ADJ ajst-6287 77 26 - - PUNCT ajst-6287 77 27 training training NOUN ajst-6287 77 28 set set NOUN ajst-6287 77 29 . . PUNCT ajst-6287 78 1 the the DET ajst-6287 78 2 classification classification NOUN ajst-6287 78 3 accuracy accuracy NOUN ajst-6287 78 4 of of ADP ajst-6287 78 5 all all DET ajst-6287 78 6 tasks task NOUN ajst-6287 78 7 is be AUX ajst-6287 78 8 averaged average VERB ajst-6287 78 9 to to PART ajst-6287 78 10 evaluate evaluate VERB ajst-6287 78 11 the the DET ajst-6287 78 12 classification classification NOUN ajst-6287 78 13 effect effect NOUN ajst-6287 78 14 under under ADP ajst-6287 78 15 the the DET ajst-6287 78 16 corresponding corresponding ADJ ajst-6287 78 17 parameters parameter NOUN ajst-6287 78 18 . . PUNCT ajst-6287 79 1 table table NOUN ajst-6287 79 2 1 1 NUM ajst-6287 79 3 . . PUNCT ajst-6287 79 4 comparison comparison NOUN ajst-6287 79 5 of of ADP ajst-6287 79 6 classification classification NOUN ajst-6287 79 7 accuracy accuracy NOUN ajst-6287 79 8 for for ADP ajst-6287 79 9 different different ADJ ajst-6287 79 10 parameters parameter NOUN ajst-6287 79 11 category category NOUN ajst-6287 79 12 number number NOUN ajst-6287 79 13 k=1 k=1 PROPN ajst-6287 80 1 k=5 k=5 PROPN ajst-6287 80 2 k=10 k=10 X ajst-6287 80 3 n=3 n=3 SYM ajst-6287 80 4 0.783 0.783 NUM ajst-6287 80 5 0.853 0.853 NUM ajst-6287 80 6 0.886 0.886 NUM ajst-6287 80 7 n=5 n=5 NUM ajst-6287 80 8 0.691 0.691 NUM ajst-6287 80 9 0.812 0.812 NUM ajst-6287 80 10 0.848 0.848 NUM ajst-6287 80 11 table table NOUN ajst-6287 80 12 1 1 NUM ajst-6287 80 13 compares compare VERB ajst-6287 80 14 the the DET ajst-6287 80 15 classification classification NOUN ajst-6287 80 16 accuracy accuracy NOUN ajst-6287 80 17 of of ADP ajst-6287 80 18 plant plant NOUN ajst-6287 80 19 leaf leaf NOUN ajst-6287 80 20 diseases disease NOUN ajst-6287 80 21 using use VERB ajst-6287 80 22 different different ADJ ajst-6287 80 23 n n CCONJ ajst-6287 80 24 - - PUNCT ajst-6287 80 25 way way NOUN ajst-6287 80 26 k k ADJ ajst-6287 80 27 - - PUNCT ajst-6287 80 28 shot shot NOUN ajst-6287 80 29 parameters parameter NOUN ajst-6287 80 30 . . PUNCT ajst-6287 81 1 it it PRON ajst-6287 81 2 can can AUX ajst-6287 81 3 be be AUX ajst-6287 81 4 seen see VERB ajst-6287 81 5 that that SCONJ ajst-6287 81 6 when when SCONJ ajst-6287 81 7 the the DET ajst-6287 81 8 k k PROPN ajst-6287 81 9 value value NOUN ajst-6287 81 10 is be AUX ajst-6287 81 11 the the DET ajst-6287 81 12 same same ADJ ajst-6287 81 13 , , PUNCT ajst-6287 81 14 smaller small ADJ ajst-6287 81 15 n n NOUN ajst-6287 81 16 value value NOUN ajst-6287 81 17 results result NOUN ajst-6287 81 18 in in ADP ajst-6287 81 19 higher high ADJ ajst-6287 81 20 classification classification NOUN ajst-6287 81 21 accuracy accuracy NOUN ajst-6287 81 22 . . PUNCT ajst-6287 82 1 this this PRON ajst-6287 82 2 is be AUX ajst-6287 82 3 because because SCONJ ajst-6287 82 4 n n PRON ajst-6287 82 5 represents represent VERB ajst-6287 82 6 the the DET ajst-6287 82 7 number number NOUN ajst-6287 82 8 to to PART ajst-6287 82 9 be be AUX ajst-6287 82 10 classified classify VERB ajst-6287 82 11 . . PUNCT ajst-6287 83 1 a a DET ajst-6287 83 2 smaller small ADJ ajst-6287 83 3 n n NOUN ajst-6287 83 4 means mean VERB ajst-6287 83 5 a a DET ajst-6287 83 6 simpler simple ADJ ajst-6287 83 7 classification classification NOUN ajst-6287 83 8 task task NOUN ajst-6287 83 9 , , PUNCT ajst-6287 83 10 which which PRON ajst-6287 83 11 corresponds correspond VERB ajst-6287 83 12 a a DET ajst-6287 83 13 larger large ADJ ajst-6287 83 14 probability probability NOUN ajst-6287 83 15 of of ADP ajst-6287 83 16 getting get VERB ajst-6287 83 17 the the DET ajst-6287 83 18 correct correct ADJ ajst-6287 83 19 classification classification NOUN ajst-6287 83 20 result result NOUN ajst-6287 83 21 . . PUNCT ajst-6287 84 1 when when SCONJ ajst-6287 84 2 n n X ajst-6287 84 3 is be AUX ajst-6287 84 4 the the DET ajst-6287 84 5 same same ADJ ajst-6287 84 6 , , PUNCT ajst-6287 84 7 a a DET ajst-6287 84 8 larger large ADJ ajst-6287 84 9 k k NOUN ajst-6287 84 10 value value NOUN ajst-6287 84 11 would would AUX ajst-6287 84 12 acquire acquire VERB ajst-6287 84 13 a a DET ajst-6287 84 14 higher high ADJ ajst-6287 84 15 classification classification NOUN ajst-6287 84 16 accuracy accuracy NOUN ajst-6287 84 17 . . PUNCT ajst-6287 85 1 this this PRON ajst-6287 85 2 is be AUX ajst-6287 85 3 because because SCONJ ajst-6287 85 4 k k PROPN ajst-6287 85 5 represents represent VERB ajst-6287 85 6 the the DET ajst-6287 85 7 number number NOUN ajst-6287 85 8 of of ADP ajst-6287 85 9 samples sample NOUN ajst-6287 85 10 in in ADP ajst-6287 85 11 each each DET ajst-6287 85 12 category category NOUN ajst-6287 85 13 . . PUNCT ajst-6287 86 1 a a DET ajst-6287 86 2 larger large ADJ ajst-6287 86 3 k k PROPN ajst-6287 86 4 value value NOUN ajst-6287 86 5 means mean VERB ajst-6287 86 6 more more ADJ ajst-6287 86 7 information information NOUN ajst-6287 86 8 in in ADP ajst-6287 86 9 each each DET ajst-6287 86 10 category category NOUN ajst-6287 86 11 . . PUNCT ajst-6287 87 1 thus thus ADV ajst-6287 87 2 the the DET ajst-6287 87 3 calculated calculate VERB ajst-6287 87 4 prototype prototype NOUN ajst-6287 87 5 features feature NOUN ajst-6287 87 6 can can AUX ajst-6287 87 7 represent represent VERB ajst-6287 87 8 the the DET ajst-6287 87 9 corresponding correspond VERB ajst-6287 87 10 category category NOUN ajst-6287 87 11 more more ADV ajst-6287 87 12 accurately accurately ADV ajst-6287 87 13 , , PUNCT ajst-6287 87 14 which which PRON ajst-6287 87 15 results result VERB ajst-6287 87 16 in in ADP ajst-6287 87 17 higher high ADJ ajst-6287 87 18 classification classification NOUN ajst-6287 87 19 accuracy accuracy NOUN ajst-6287 87 20 . . PUNCT ajst-6287 88 1 4 4 X ajst-6287 88 2 . . X ajst-6287 88 3 summary summary NOUN ajst-6287 88 4 in in ADP ajst-6287 88 5 this this DET ajst-6287 88 6 paper paper NOUN ajst-6287 88 7 , , PUNCT ajst-6287 88 8 the the DET ajst-6287 88 9 plant plant NOUN ajst-6287 88 10 leaf leaf NOUN ajst-6287 88 11 disease disease NOUN ajst-6287 88 12 classification classification NOUN ajst-6287 88 13 was be AUX ajst-6287 88 14 studied study VERB ajst-6287 88 15 based base VERB ajst-6287 88 16 on on ADP ajst-6287 88 17 few few ADJ ajst-6287 88 18 - - PUNCT ajst-6287 88 19 shot shot NOUN ajst-6287 88 20 learning learning NOUN ajst-6287 88 21 method method NOUN ajst-6287 88 22 . . PUNCT ajst-6287 89 1 n n CCONJ ajst-6287 89 2 - - PUNCT ajst-6287 89 3 way way NOUN ajst-6287 89 4 k k NOUN ajst-6287 89 5 - - PUNCT ajst-6287 89 6 shot shot ADJ ajst-6287 89 7 tasks task NOUN ajst-6287 89 8 were be AUX ajst-6287 89 9 constructed construct VERB ajst-6287 89 10 for for ADP ajst-6287 89 11 model model NOUN ajst-6287 89 12 training training NOUN ajst-6287 89 13 through through ADP ajst-6287 89 14 metalearning metalearning NOUN ajst-6287 89 15 , , PUNCT ajst-6287 89 16 and and CCONJ ajst-6287 89 17 the the DET ajst-6287 89 18 category category NOUN ajst-6287 89 19 of of ADP ajst-6287 89 20 each each DET ajst-6287 89 21 sample sample NOUN ajst-6287 89 22 in in ADP ajst-6287 89 23 the the DET ajst-6287 89 24 meta meta ADJ ajst-6287 89 25 - - PUNCT ajst-6287 89 26 test test NOUN ajst-6287 89 27 dataset dataset NOUN ajst-6287 89 28 is be AUX ajst-6287 89 29 predicted predict VERB ajst-6287 89 30 by by ADP ajst-6287 89 31 the the DET ajst-6287 89 32 prototype prototype NOUN ajst-6287 89 33 features feature VERB ajst-6287 89 34 . . PUNCT ajst-6287 90 1 some some DET ajst-6287 90 2 key key ADJ ajst-6287 90 3 features feature NOUN ajst-6287 90 4 affecting affect VERB ajst-6287 90 5 the the DET ajst-6287 90 6 classification classification NOUN ajst-6287 90 7 accuracy accuracy NOUN ajst-6287 90 8 are be AUX ajst-6287 90 9 revealed reveal VERB ajst-6287 90 10 by by ADP ajst-6287 90 11 changing change VERB ajst-6287 90 12 the the DET ajst-6287 90 13 parameter parameter NOUN ajst-6287 90 14 configuration configuration NOUN ajst-6287 90 15 of of ADP ajst-6287 90 16 the the DET ajst-6287 90 17 learning learning NOUN ajst-6287 90 18 process process NOUN ajst-6287 90 19 . . PUNCT ajst-6287 91 1 84 84 NUM ajst-6287 91 2 future future ADJ ajst-6287 91 3 studies study NOUN ajst-6287 91 4 will will AUX ajst-6287 91 5 continue continue VERB ajst-6287 91 6 to to PART ajst-6287 91 7 explore explore VERB ajst-6287 91 8 the the DET ajst-6287 91 9 application application NOUN ajst-6287 91 10 of of ADP ajst-6287 91 11 few few ADJ ajst-6287 91 12 - - PUNCT ajst-6287 91 13 shot shot NOUN ajst-6287 91 14 learning learning NOUN ajst-6287 91 15 method method NOUN ajst-6287 91 16 in in ADP ajst-6287 91 17 intelligent intelligent ADJ ajst-6287 91 18 agriculture agriculture NOUN ajst-6287 91 19 . . PUNCT ajst-6287 92 1 acknowledgment acknowledgment NOUN ajst-6287 92 2 this this DET ajst-6287 92 3 work work NOUN ajst-6287 92 4 was be AUX ajst-6287 92 5 supported support VERB ajst-6287 92 6 by by ADP ajst-6287 92 7 the the DET ajst-6287 92 8 tai'an tai'an PROPN ajst-6287 92 9 science science NOUN ajst-6287 92 10 and and CCONJ ajst-6287 92 11 technology technology NOUN ajst-6287 92 12 innovation innovation NOUN ajst-6287 92 13 development development NOUN ajst-6287 92 14 project project NOUN ajst-6287 92 15 under under ADP ajst-6287 92 16 grant grant NOUN ajst-6287 92 17 no.2021ns097 no.2021ns097 PROPN ajst-6287 92 18 . . PUNCT ajst-6287 93 1 references reference NOUN ajst-6287 93 2 [ [ X ajst-6287 93 3 1 1 X ajst-6287 93 4 ] ] PUNCT ajst-6287 93 5 ferentinos ferentinos PROPN ajst-6287 93 6 k k PROPN ajst-6287 93 7 p.deep p.deep PROPN ajst-6287 93 8 learning learning NOUN ajst-6287 93 9 models model NOUN ajst-6287 93 10 for for ADP ajst-6287 93 11 plant plant NOUN ajst-6287 93 12 disease disease NOUN ajst-6287 93 13 detection detection NOUN ajst-6287 93 14 and and CCONJ ajst-6287 93 15 diagnosis[j].computers diagnosis[j].computer NOUN ajst-6287 93 16 and and CCONJ ajst-6287 93 17 electronics electronic NOUN ajst-6287 93 18 in in ADP ajst-6287 93 19 agriculture agriculture NOUN ajst-6287 93 20 , , PUNCT ajst-6287 93 21 2018 2018 NUM ajst-6287 93 22 , , PUNCT ajst-6287 93 23 145 145 NUM ajst-6287 93 24 : : PUNCT ajst-6287 93 25 311 311 NUM ajst-6287 93 26 - - SYM ajst-6287 93 27 318 318 NUM ajst-6287 93 28 . . PUNCT ajst-6287 94 1 [ [ X ajst-6287 94 2 2 2 NUM ajst-6287 94 3 ] ] X ajst-6287 94 4 zhou zhou PROPN ajst-6287 94 5 c c X ajst-6287 94 6 , , PUNCT ajst-6287 94 7 zhou zhou PROPN ajst-6287 94 8 s s PROPN ajst-6287 94 9 , , PUNCT ajst-6287 94 10 xing xing PROPN ajst-6287 94 11 j j PROPN ajst-6287 94 12 , , PUNCT ajst-6287 94 13 et et PROPN ajst-6287 94 14 al al PROPN ajst-6287 94 15 . . PUNCT ajst-6287 94 16 tomato tomato NOUN ajst-6287 94 17 leaf leaf NOUN ajst-6287 94 18 disease disease NOUN ajst-6287 94 19 identification identification NOUN ajst-6287 94 20 by by ADP ajst-6287 94 21 restructured restructure VERB ajst-6287 94 22 deep deep ADJ ajst-6287 94 23 residual residual ADJ ajst-6287 94 24 dense dense ADJ ajst-6287 94 25 network[j network[j NOUN ajst-6287 94 26 ] ] PUNCT ajst-6287 94 27 . . PUNCT ajst-6287 95 1 ieee ieee NOUN ajst-6287 95 2 access access NOUN ajst-6287 95 3 , , PUNCT ajst-6287 95 4 2021 2021 NUM ajst-6287 95 5 , , PUNCT ajst-6287 95 6 9 9 NUM ajst-6287 95 7 : : SYM ajst-6287 95 8 28822 28822 NUM ajst-6287 95 9 - - SYM ajst-6287 95 10 28831 28831 NUM ajst-6287 95 11 . . PUNCT ajst-6287 96 1 [ [ X ajst-6287 96 2 3 3 X ajst-6287 96 3 ] ] X ajst-6287 96 4 chen chen PROPN ajst-6287 96 5 j j PROPN ajst-6287 96 6 , , PUNCT ajst-6287 96 7 chen chen PROPN ajst-6287 96 8 j j PROPN ajst-6287 96 9 , , PUNCT ajst-6287 96 10 zhang zhang PROPN ajst-6287 96 11 d d PROPN ajst-6287 96 12 , , PUNCT ajst-6287 96 13 et et PROPN ajst-6287 96 14 al al PROPN ajst-6287 96 15 . . PUNCT ajst-6287 97 1 using use VERB ajst-6287 97 2 deep deep ADJ ajst-6287 97 3 transfer transfer NOUN ajst-6287 97 4 learning learning NOUN ajst-6287 97 5 for for ADP ajst-6287 97 6 image image NOUN ajst-6287 97 7 - - PUNCT ajst-6287 97 8 based base VERB ajst-6287 97 9 plant plant NOUN ajst-6287 97 10 disease disease NOUN ajst-6287 97 11 identification[j identification[j PROPN ajst-6287 97 12 ] ] PUNCT ajst-6287 97 13 . . PUNCT ajst-6287 98 1 computers computer NOUN ajst-6287 98 2 and and CCONJ ajst-6287 98 3 electronics electronic NOUN ajst-6287 98 4 in in ADP ajst-6287 98 5 agriculture agriculture NOUN ajst-6287 98 6 , , PUNCT ajst-6287 98 7 2020 2020 NUM ajst-6287 98 8 , , PUNCT ajst-6287 98 9 173 173 NUM ajst-6287 98 10 : : SYM ajst-6287 98 11 105393 105393 NUM ajst-6287 98 12 . . PUNCT ajst-6287 99 1 [ [ X ajst-6287 99 2 4 4 X ajst-6287 99 3 ] ] PUNCT ajst-6287 99 4 szegedy szegedy VERB ajst-6287 99 5 c c NOUN ajst-6287 99 6 , , PUNCT ajst-6287 99 7 vanhoucke vanhoucke NOUN ajst-6287 99 8 v v NOUN ajst-6287 99 9 , , PUNCT ajst-6287 99 10 ioffe ioffe PROPN ajst-6287 99 11 s s PART ajst-6287 99 12 , , PUNCT ajst-6287 99 13 et et PROPN ajst-6287 99 14 al al PROPN ajst-6287 99 15 . . PUNCT ajst-6287 100 1 rethinking rethink VERB ajst-6287 100 2 the the DET ajst-6287 100 3 inception inception ADJ ajst-6287 100 4 architecture architecture NOUN ajst-6287 100 5 for for ADP ajst-6287 100 6 computer computer NOUN ajst-6287 100 7 vision[c]//proceedings vision[c]//proceeding NOUN ajst-6287 100 8 of of ADP ajst-6287 100 9 the the DET ajst-6287 100 10 ieee ieee NOUN ajst-6287 100 11 conference conference NOUN ajst-6287 100 12 on on ADP ajst-6287 100 13 computer computer NOUN ajst-6287 100 14 vision vision NOUN ajst-6287 100 15 and and CCONJ ajst-6287 100 16 pattern pattern NOUN ajst-6287 100 17 recognition recognition NOUN ajst-6287 100 18 . . PUNCT ajst-6287 101 1 2016 2016 NUM ajst-6287 101 2 : : PUNCT ajst-6287 101 3 2818 2818 NUM ajst-6287 101 4 - - SYM ajst-6287 101 5 2826 2826 NUM ajst-6287 101 6 . . PUNCT ajst-6287 102 1 [ [ X ajst-6287 102 2 5 5 NUM ajst-6287 102 3 ] ] X ajst-6287 102 4 snell snell PROPN ajst-6287 102 5 j j PROPN ajst-6287 102 6 , , PUNCT ajst-6287 102 7 swersky swersky PROPN ajst-6287 102 8 k k PROPN ajst-6287 102 9 , , PUNCT ajst-6287 102 10 zemel zemel NOUN ajst-6287 102 11 r r PROPN ajst-6287 102 12 s. s. PROPN ajst-6287 102 13 prototypical prototypical ADJ ajst-6287 102 14 networks network NOUN ajst-6287 102 15 for for ADP ajst-6287 102 16 fewshot fewshot ADJ ajst-6287 102 17 learning[j learning[j NOUN ajst-6287 102 18 ] ] PUNCT ajst-6287 102 19 . . PUNCT ajst-6287 103 1 arxiv arxiv PROPN ajst-6287 103 2 preprint preprint PROPN ajst-6287 103 3 arxiv:1703.05175 arxiv:1703.05175 PROPN ajst-6287 103 4 , , PUNCT ajst-6287 103 5 2017 2017 NUM ajst-6287 103 6 . . PUNCT ajst-6287 104 1 [ [ X ajst-6287 104 2 6 6 NUM ajst-6287 104 3 ] ] PUNCT ajst-6287 104 4 hughes hughes PROPN ajst-6287 104 5 d d PROPN ajst-6287 104 6 , , PUNCT ajst-6287 104 7 salathΓ© salathΓ© ADJ ajst-6287 104 8 m. m. NOUN ajst-6287 104 9 an an DET ajst-6287 104 10 open open ADJ ajst-6287 104 11 access access NOUN ajst-6287 104 12 repository repository NOUN ajst-6287 104 13 of of ADP ajst-6287 104 14 images image NOUN ajst-6287 104 15 on on ADP ajst-6287 104 16 plant plant NOUN ajst-6287 104 17 health health NOUN ajst-6287 104 18 to to PART ajst-6287 104 19 enable enable VERB ajst-6287 104 20 the the DET ajst-6287 104 21 development development NOUN ajst-6287 104 22 of of ADP ajst-6287 104 23 mobile mobile ADJ ajst-6287 104 24 disease disease NOUN ajst-6287 104 25 diagnostics[j diagnostics[j PROPN ajst-6287 104 26 ] ] PUNCT ajst-6287 104 27 . . PUNCT ajst-6287 105 1 arxiv arxiv PROPN ajst-6287 105 2 preprint preprint VERB ajst-6287 105 3 arxiv:1511.08060 arxiv:1511.08060 NOUN ajst-6287 105 4 , , PUNCT ajst-6287 105 5 2015 2015 NUM ajst-6287 105 6 . . PUNCT