$30 off During Our Annual Pro Sale. View Details »
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
文献紹介:CMMC-BDRC Solution to the NLP-TEA-2018 Chi...
Search
Atsushi
September 28, 2018
0
130
文献紹介:CMMC-BDRC Solution to the NLP-TEA-2018 Chinese Grammatical Error Diagnosis Task
2018年9月28日 文献紹介
長岡技術科学大学
自然言語処理研究室
Atsushi
September 28, 2018
Tweet
Share
More Decks by Atsushi
See All by Atsushi
文献紹介:Automated Evaluation of Out-of-Context Errors
atsumikan
0
98
文献紹介:Correction of OCR Word Segmentation Errors in Articles from the ACL Collection through Neural Machine Translation Methods
atsumikan
0
160
文献紹介:Auxiliary Objectives for Neural Error Detection Models
atsumikan
0
92
文献紹介:Wronging a Right: Generating Better Errors to Improve Grammatical Error Detection
atsumikan
0
120
文献紹介:Low-resource OCR error detection and correction in French Clinical Texts
atsumikan
0
130
文献紹介 : Fluency Boost Learning and Inference for Neural Grammatical Error Correction
atsumikan
0
180
文献紹介:語彙の概念化と Wikipediaを用いた英字略語の意味推定方法
atsumikan
0
150
文献紹介:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
atsumikan
0
130
文献紹介: Automatic Annotation and Evaluation of Error Types for Grammatical Error Correction
atsumikan
0
170
Featured
See All Featured
Understanding Cognitive Biases in Performance Measurement
bluesmoon
32
2.8k
A designer walks into a library…
pauljervisheath
210
24k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
35
3.3k
Java REST API Framework Comparison - PWX 2021
mraible
34
9k
4 Signs Your Business is Dying
shpigford
186
22k
Building a Scalable Design System with Sketch
lauravandoore
463
34k
The Art of Programming - Codeland 2020
erikaheidi
56
14k
Put a Button on it: Removing Barriers to Going Fast.
kastner
60
4.1k
Optimising Largest Contentful Paint
csswizardry
37
3.5k
It's Worth the Effort
3n
187
29k
The Straight Up "How To Draw Better" Workshop
denniskardys
239
140k
Build The Right Thing And Hit Your Dates
maggiecrowley
38
3k
Transcript
$..$#%3$4PMVUJPOUPUIF/-1 5&"$IJOFTF(SBNNBUJDBM&SSPS %JBHOPTJT5BTL Ԭٕज़Պֶେֶࣗવݴޠॲཧݚڀࣨ ੁ३ࢤ จݙհ ݄ :POHXFJ;IBOH 2JOBO)V 'BOH-JV
BOE:VFHVP(V 1SPDFFEJOHTPGUIFUI8PSLTIPQPO/BUVSBM-BOHVBHF1SPDFTTJOH 5FDIOJRVFTGPS&EVDBUJPOBM"QQMJDBUJPOT QBHFTr.FMCPVSOF "VTUSBMJB +VMZ 1
"CTUSBDU w தࠃޠʹ͓͚Δจ๏ޡΓగਖ਼ͷλεΫʹऔΓΉ w طଘͷ#J-45.Λ༻ͨ͠ϞσϧͰτϨʔχϯά σʔλ͕গͳ͍ͷͰ֦ுΛߦͬͨ w ݕग़λεΫͰ࠷ߴ͍'είΞΛ֫ಘ w ฏۉͰ൪ʹߴ͍'είΞΛ֫ಘ
!2
*OUSPEVDUJPO w தࠃͷൃలʹ͍தࠃޠͷֶशऀ͕૿Ճ w ӳޠΑΓதࠃޠͷจ๏͕ෳࡶ w தࠃޠΛޠͱ͠ͳֶ͍शऀͷशಘ͕ࠔ !3
*OUSPEVDUJPO w $(&%ɿதࠃޠͷจ๏ޡΓΛగਖ਼͢ΔλεΫ w ޡΓͷҐஔɺछྨɺగਖ਼ީิΛݟ͚ͭΔ w ޡΓͷछྨ !4 w 4ͱ.࠷େͭ·Ͱͷగਖ਼ީิΛఏҊͰ͖Δ
ͳ୯ޠʢ3ʣ ୯ޠͷܽམʢ.ʣ ୯ޠͷબޡΓʢ4ʣ ୯ޠͷॱ൪ͷޡΓʢ8ʣ
*OUSPEVDUJPO w ݕग़λεΫɿจষޡ͍ͬͯΔͷ͔ʁ w ࣝผλεΫɿͲͷޡΓͷछྨ͕ຒΊࠐ·Ε͍ͯΔ͔ʁ w ҐஔλεΫɿΤϥʔҐஔ͕ൃੜ͍ͯ͠ΔൣғͲ͔͜ʁ w గਖ਼λεΫɿగਖ਼ޙͷ୯ޠԿͰ͋Δ͔ʁ !5
ݕग़Ϟσϧ w )*5UFBN ;IFOHFUBM ٴͼ"MJCBCBUFBN :BOHFUBM ͷख๏Λ࠾༻ !6
గਖ਼Ϟσϧ w ςΩετྨͷλεΫͱΈͳ͠గਖ਼ w ΧςΰϦɿਖ਼͍͠୯ޠ w จॻɿޡͬͨ୯ޠͱͦͷલޙͷ୯ޠ !7
τϨʔχϯάσʔλͷ֦ுͷखॱ ޡΓͷϧʔϧΛநग़͢Δ w (&$%τϨʔχϯάηοτ͔Βநग़ w ωΠςΟϒεϐʔΧʔ͕ޡΓ͍͢୯ޠ͔Βநग़ σʔλੜ w ϧʔϧʹै͍ਖ਼͍͠จΛޡͬͨจʹม !8
τϨʔχϯάσʔλͷ֦ுͷखॱ ޡΓͷϧʔϧΛநग़͢Δ w (&$%τϨʔχϯάηοτ͔Βநग़ w ωΠςΟϒεϐʔΧʔ͕ޡΓ͍͢୯ޠ͔Βநग़ σʔλੜ w ϧʔϧʹै͍ਖ਼͍͠จΛޡͬͨจʹม !9
ϧʔϧͷநग़ w $(&%τϨʔχϯάηοτ͔ΒޡΓͷϧʔϧΛநग़ w ϧʔϧΛநग़͢ΔͨΊͷखॱ ޡͬͨจͱਖ਼͍͠จͷ͕ҧ͏จষΛআڈ ޡͬͨจͱਖ਼͍͠จΛׂ ޡͬͨ୯ޠͱͦͷલޙͷ୯ޠ͔ΒϧʔϧΛ࡞ !10
ϧʔϧͷநग़ w ಘΒΕΔϧʔϧ w bछྨޡͬͨจࣈલͷ୯ޠਖ਼͍͠จࣈޙͷ୯ޠ` w b4ڭ䟙ሇࢠڢ䟙ਓ` w b4ڭ䟙ࢠڢ䟙` !11
ϧʔϧͷநग़ w ωΠςΟϒεϐʔΧʔͷޡΓ͔Βͷϧʔϧநग़ w ωΠςΟϒεϐʔΧʔ͕ޡΓ͍͢୯ޠϦετ͕ଘࡏ w ͦͷલޙʹग़ݱ͢Δ୯ޠهࡌ͞Ε͍ͯͳ͍ͷͰ ϧʔϧΛநग़Ͱ͖ͳ͍ w ίʔύε͔ΒޡΓ͍͢୯ޠͷલޙͷ୯ޠΛநग़
!12
ϧʔϧͷநग़ w લॲཧ $(&%σʔληοτΛมͯ͠؆ૉԽ͢Δ ୯ޠͷׂΛߦ͏ จষΛϑΟϧλϦϯά͢Δ wࣈͷׂ߹จࣈ͕গͳ͍ͷ wதࠃޠֶशऀ͕༻͠ͳ͍୯ޠΛؚΉͷ !13
τϨʔχϯάσʔλͷ֦ுͷखॱ ޡΓͷϧʔϧΛநग़͢Δ w (&$%τϨʔχϯάηοτ͔Βநग़ w ωΠςΟϒεϐʔΧʔ͕ޡΓ͍͢୯ޠ͔Βநग़ σʔλੜ w ϧʔϧʹै͍ਖ਼͍͠จΛޡͬͨจʹม !14
࣮ݧ݅ w ϧʔϧ w աڈͷ$(&%τϨʔχϯάσʔλͱςετσʔλɿ ݸ w ڭՊॻͳͲͷίʔύεɿ ݸ
w ੜͨ͠Τϥʔจɿ จ w ׂɿ܇࿅σʔλ w ׂɿݕূͷͨΊͷσʔλ !15
!16
࣮ݧ݁ՌʢධՁσʔληοτʣ !17
$PODMVTJPO w ϞσϧͷτϨʔχϯάηοτΛ֦ு͢ΔͨΊʹ ΑΓଟ͘ͷΤϥʔจΛੜͨ͠ w ฏۉϥϯΩϯάͰ൪ʹߴ͍'είΞΛ֫ಘ w σʔλ֦ுΞϧΰϦζϜͷ༗ޮੑΛূ໌ w ࠓޙͷ࡞ۀ
w Τϥʔݕग़ͱमਖ਼ͰΑΓଟ͘ͷจ๏తͳಛΛ͏ w ୈೋݴޠͱͯ͠தࠃޠΛڭ͑ΔࡍͷܦݧΛ׆͔͢ !18