Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
文献紹介:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
Search
Atsushi
June 27, 2018
0
110
文献紹介:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
2018年6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
June 27, 2018
Tweet
Share
More Decks by Atsushi
See All by Atsushi
文献紹介:Automated Evaluation of Out-of-Context Errors
atsumikan
0
76
文献紹介:Correction of OCR Word Segmentation Errors in Articles from the ACL Collection through Neural Machine Translation Methods
atsumikan
0
130
文献紹介:Auxiliary Objectives for Neural Error Detection Models
atsumikan
0
66
文献紹介:Wronging a Right: Generating Better Errors to Improve Grammatical Error Detection
atsumikan
0
93
文献紹介:Low-resource OCR error detection and correction in French Clinical Texts
atsumikan
0
75
文献紹介:CMMC-BDRC Solution to the NLP-TEA-2018 Chinese Grammatical Error Diagnosis Task
atsumikan
0
93
文献紹介 : Fluency Boost Learning and Inference for Neural Grammatical Error Correction
atsumikan
0
150
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
120
文献紹介: Automatic Annotation and Evaluation of Error Types for Grammatical Error Correction
atsumikan
0
140
Featured
See All Featured
Typedesign – Prime Four
hannesfritz
36
2.1k
The Power of CSS Pseudo Elements
geoffreycrofte
60
5k
Rebuilding a faster, lazier Slack
samanthasiow
73
8.2k
A Philosophy of Restraint
colly
197
16k
GitHub's CSS Performance
jonrohan
1025
450k
Code Review Best Practice
trishagee
55
15k
XXLCSS - How to scale CSS and keep your sanity
sugarenia
241
1.2M
Fight the Zombie Pattern Library - RWD Summit 2016
marcelosomers
227
16k
StorybookのUI Testing Handbookを読んだ
zakiyama
13
4.6k
Mobile First: as difficult as doing things right
swwweet
216
8.6k
VelocityConf: Rendering Performance Case Studies
addyosmani
320
23k
Building Effective Engineering Teams - LeadDev
addyosmani
28
1.8k
Transcript
5IF&⒎FDUPG&SSPS3BUF JO"SUJpDJBMMZ(FOFSBUFE%BUB GPS"VUPNBUJD1SFQPTJUJPO BOE%FUFSNJOFS$PSSFDUJPO Ԭٕज़Պֶେֶࣗવݴޠॲཧݚڀࣨ ੁ३ࢤ จݙհ ݄ 'SBTFS#PXFOBOE+PO%FIEBSJBOE+PTFGWBO(FOBCJUI 1SPDFFEJOHTPGUIFSE8PSLTIPQPO/PJTZ6TFSHFOFSBUFE5FYU
QBHFTr$PQFOIBHFO %FONBSL 4FQUFNCFS
"CTUSBDU w લஔࢺͱႊఆࢺΛగਖ਼͢ΔχϡʔϥϧγεςϜʹ͓͍ͯɺ ܇࿅σʔλͱςετσʔλͷؒͷΤϥʔͷׂ߹ͱछྨͷϛ εϚονͷӨڹΛௐࠪ w ΤϥʔλΠϓΛΈ߹Θֶͤͯश͢Δ͜ͱՄೳͰ͋Δ ͕ɺ࠷ྑͷ݁ՌΛಘΔͨΊʹɺલஔࢺͱܾఆࢺͰɺ܇࿅ σʔλʹҙਤతʹೖΕΔ͖Τϥʔͷྔ͕ҟͳΔ !2
*OUSPEVDUJPO w (&$Ͱχϡʔϥϧ༁Ϟσϧɺಛʹ4FR4FRͷ༁͕༏ ҐΛอ͍ͬͯΔ w ͔͠͠ɺχϡʔϥϧωοτϫʔΫʹ࣍ͷΑ͏ͳ՝͕͋Δ w େྔͷ܇࿅σʔλΛඞཁͱ͢Δ͕Ξ ϊςʔγϣϯ͞Εͨσʔ λ͕গͳ͍
w ҟͳΔςετηοτʹΘͨͬͯҰൠԽ͞Ε͍ͯͳ͍ w τϨʔχϯάσʔλͷΤϥʔͷʹࠨӈ͞Ε͍͢ !3
*OUSPEVDUJPO w ܇࿅ίʔύεʹ͓͚ΔΤϥʔͷྔ͕గਖ਼ʹٴ΅͢Өڹʹͭ ͍ͯௐࠪ͢Δ w 3P[PWTLBZBFUBM ͷz&SSPS*OqBUJPOzͷݕূ w 3FDBMMΛվળ͢ΔͨΊʹɺඞཁҎ্ʹଟ͘ͷΤϥʔΛτ
ϨʔχϯάσʔλʹՃ͢Δ͜ͱͰɺϞσϧ͕ҰൠԽ͢ Δͷʹཱͭ !4
%BUB w ਖ਼͍͠σʔλ w 8.5/FXT$SBXMίʔύε w ܇࿅ɿ จ w
։ൃɿ จ w ϊΠζͷੜ w $BIJMMFUBM ͷख๏ʹج͍ͮͯ࡞ w -BOHͷగਖ਼σʔλ w 8JLJQFEJBͷగਖ਼σʔλ !5
%BUB w ΤϥʔΛˋɺˋɺˋɺˋؚΉτϨʔχϯάσʔλΛ࡞ w લஔࢺͷΈͷΤϥʔ w 8JLJQFEJB -BOHίʔύε w ܾఆࢺͷΈͷΤϥʔ
w -BOHίʔύε w લஔࢺٴͼܾఆࢺͷΤϥʔ w 8JLJQFEJB -BOHίʔύε !6
%BUB w ΤϥʔΛˋɺˋɺˋɺˋؚΉςετσʔλΛ࡞ w -BOHίʔύε w લஔࢺ ܾఆࢺ ྆ํͷޡΓΛؚΉจɿ֤จ w
/6$-&ίʔύε w લஔࢺͷޡΓΛؚΉจɿจ w ܾఆࢺͷޡΓΛؚΉจɿจ w ྆ํͷޡΓΛؚΉจɿจ !7
&YQFSJNFOUT w ख๏ɿ:VBOBOE#SJTDPF w ධՁख๏ɿ(-&6 !8
3FTVMUT w ޡࠩΛؚΉςετσʔλʹ͓͍ͯɺߴ͍είΞΛ࣋ͭ w จষ͕గਖ਼͞ΕΔඞཁ͕ͳ͍͜ͱΛϞσϧ͕ೝࣝ͢Δ͜ͱ͕Ͱ͖͍ͯΔͨΊ w z&SSPS*OqBUJPOz͕ػೳ͍ͯ͠Δ !9
!10 3FTVMUT wલஔࢺͷ߹ͱಉ༷ͳ݁Ռ͕ಘΒΕͨ wz&SSPS*OqBUJPOzػೳ͍ͯ͠ͳ͍ wܾఆࢺΛमਖ਼͢ΔγεςϜ͕ɺલஔࢺΛमਖ਼͢ΔγεςϜͱҟͳΔੑ࣭ Λ༗͢Δ͜ͱΛࣔ͢
!11 3FTVMUT wෳ߹ͷΤϥʔλΠϓͰɺ୯ҰͷΤϥʔΑΓείΞ͕͘ͳΔ wಛʹ/6$-&ͷείΞ͕ɺ୯ҰͷΤϥʔͷ߹ΑΓେ͖͍͘είΞͱͳ͍ͬͯΔɻ͜Ε ɺυϝΠϯؒͰҰൠԽͰ͖ͳ͔ͬͨ͜ͱΛࣔ͢ wz&SSPS*OqBUJPOzػೳ͍ͯ͠ͳ͍
3FTVMUT w z&SSPS*OqBUJPOzʹٙ೦ w ͦΕ͕ޡࠩͷछྨʹґଘ͠ɺલஔࢺ͕z&SSPS*OqBUJPOzͷ༻ Λྭ͢ΔछྨͰ͋ͬͨ߹ɺͦΕগͳ͘ͱෳ߹Ϟσϧ ʹଘࡏ͖͢Ͱ͋Δɻ w ෳ߹Ϟσϧʹ͓͍ͯɺˋͷ/6$-&ςετηοτͰˋΤ ϥʔΛؚΉͷτϨʔχϯάσʔλͷํ͕ˋͷͷΑΓ༏
Ε͍ͯΔɻ w ͜Εɺಋೖ͞ΕͨΤϥʔͷྔΛݮΒ͢͜ͱͰվળ͞ΕΔ ͜ͱΛ͍ࣔࠦͯ͠Δɻ !12
$PODMVTJPO w ͭͷಛఆͷΤϥʔλΠϓʹऔΓΉ͜ͱͰɺҙਤతʹੜ ͨ͠ΤϥʔΛؚΊΔΤϥʔͷछྨΛબ͢ΔΛ ໌Β͔ʹ͢Δ͜ͱΛࢦͨ͠ɻ w ಛʹલஔࢺͷ߹ʹz&SSPS*OqBUJPOzͷͱ͓Γͱͳͬͨ w τϨʔχϯάσʔλʹΤϥʔΛଟؚ͘ΊΔͱ(-&6εί Ξ͕ߴ͘ͳΔ͜ͱΛࣔ͢
w ͜ͷӨڹෳ߹ϞσϧͰ֬ೝ͞Εͳ͔ͬͨ !13