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:Low-resource OCR error detection and correction in French Clinical Texts
Search
Atsushi
October 31, 2018
0
75
文献紹介:Low-resource OCR error detection and correction in French Clinical Texts
2018/10/31 文献紹介
長岡技術科学大学
自然言語処理研究室
Atsushi
October 31, 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: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:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
atsumikan
0
110
文献紹介: Automatic Annotation and Evaluation of Error Types for Grammatical Error Correction
atsumikan
0
140
Featured
See All Featured
Docker and Python
trallard
34
2.7k
In The Pink: A Labor of Love
frogandcode
138
21k
Typedesign – Prime Four
hannesfritz
36
2.1k
Automating Front-end Workflow
addyosmani
1356
200k
The World Runs on Bad Software
bkeepers
PRO
61
6.7k
How To Stay Up To Date on Web Technology
chriscoyier
782
250k
Fontdeck: Realign not Redesign
paulrobertlloyd
76
4.9k
Teambox: Starting and Learning
jrom
128
8.4k
Git: the NoSQL Database
bkeepers
PRO
422
63k
Scaling GitHub
holman
457
140k
Done Done
chrislema
178
15k
JavaScript: Past, Present, and Future - NDC Porto 2020
reverentgeek
40
4.4k
Transcript
-PXSFTPVSDF0$3FSSPSEFUFDUJPOBOE DPSSFDUJPOJO'SFODI$MJOJDBM5FYUT Ԭٕज़Պֶେֶࣗવݴޠॲཧݚڀࣨ ੁ३ࢤ จݙհ ݄ &WB%`IPOEU $ZSJM(SPVJO #SJHJUUF(SBV 1SPDFFEJOHTPGUIF4FWFOUI*OUFSOBUJPOBM8PSLTIPQPO)FBMUI5FYU
.JOJOHBOE*OGPSNBUJPO"OBMZTJT -06)* QBHFTr "VTUJO 59 /PWFNCFS 1
"CTUSBDU w ϑϥϯεͷҩྍࢿྉʹର͢Δ0$3ޡΓͷݕग़ͱగਖ਼ w গͳ͍܇࿅σʔλͷΈΛ༻͢Δํ๏ΛఏҊ w ܇࿅࣌ʹϊΠζΛೖ͢ΔͭͷҟͳΔํ๏Λൺֱ w ˋͷਫ਼Λୡ !2
w ҩྍॻྨͷ͏ͪݹ͍ױऀͷΧϧςͷଟ͘ࢴͰอ w ҩྍݚڀͷͨΊͷوॏͳใݯ w 0$3ʹΑͬͯར༻͍͢͠σΟδλϧܗࣜʹม͍ͨ͠ w طଘͷ0$3ޡΓగਖ਼ϞσϧΛ༻Ͱ͖ͳ͍ *OUSPEVDUJPO !3
w طଘͷ0$3ޡΓగਖ਼ϞσϧΛ༻ͨ͠߹ͷ w େྔʹίʔύεΛ༻ҙͰ͖ͳ͍ w ίʔύεதͷױऀΛಛఆͰ͖Δใ͕আ͞Ε͍ͯΔ w ࢴͷྼԽҹٕ๏ɺςΩετߏʹ͕ࠩ͋Δ *OUSPEVDUJPO !4
*OUSPEVDUJPO w طଘͷ0$3ޡΓͷగਖ਼ɿ େྔͷ܇࿅σʔλΛඞཁͱ͢ΔͨΊ༻Ͱ͖ͳ͍ w υϝΠϯಛ༗ͷදతͳςΩετίʔύεͷΈΛ ඞཁͱ͢ΔγεςϜΛఏҊ !5
!6
.PEFM w ܇࿅࣌ʹͲͪΒ͔ͷख๏ͰϊΠζΛ༩͑Δ w 3BOEPNHFOFSBUJPO SBOEPN/PJTF w ޡΓͷੜͱͦͷछྨΛϥϯμϜʹܾΊΔ w
จࣈ·ͨจࣈͷআɺૠೖɺஔ w *OTFSUJPOPGDIBSBDUFSDPOGVTJPOT DPOGVTJPO1BJS w ޡΓ͍͢จࣈྻͷରΛ༻͍Δ !7
*OTFSUJPOPGDIBSBDUFSDPOGVTJPOT !8
w ϑϥϯεͷױऀΧϧςͷσʔληοτ w ױऀɿ ਓ w ॻྨɿ ຕ w ࢢൢͷ0$3ΤϯδϯͰσΟδλϧԽ
w ױऀΛࣝผͰ͖ΔใΛλάʹஔ͖͑Δ ྫʣl.BZTU zˠl%"5&z $PSQVT !9
w ίʔύεͷจॻͷ࣭Λ007ͷׂ߹ʹΑͬͯධՁ w 007ɿ6OJpFE.FEJDBM-FYJDPOGPS'SFODI ;XFJHFOCBVNFUBM ʹؚ·Εͳ͍ޠͱͨ͠ $PSQVT !10
5SBJOJOHTFU&WBMVBUJPOTFU w 5SBJOJOHTFU w l&YDFMMFOUzͷจॻ w &WBMVBUJPOTFU w l&YDFMMFOUzͱz(PPEzϥϯΫͷॻྨ͔Β ϥϯμϜʹநग़ͨ͠ͷจॻΛख࡞ۀͰగਖ਼ͨ͠ͷ
!11
$PSQVT !12
&YQFSJNFOUT w #BTFMJOF ςΩετΛτʔΫϯԽ͢Δ আɺૠೖɺஔʹΑΔ࠷ฤूڑٴͼ ޙํͷ୯ޠͱͷ݁߹ʹΑͬͯਖ਼͍͠୯ޠΛٻΊΔ ฤूૢ࡞ʹΑͬͯϥϯΫ͚Λߦ͍
࠷ߴ͍ϥϯΫͷͷΛղͱ͢Δ !13
3FTVMU !14 ख๏ "DDVSBDZ ϕʔεϥΠϯ SBOEPN/PJTF DPOGVTJPO1BJS
3FTVMU !15 w ૠೖޡΓͷగਖ਼͕ࠔ w ޡΓͷେ෦͕୯ޠʹϥϯμϜʹૠೖ͞Εۭͨന w จॻͰۭന͕ଟΘΕ͍ͯΔͨΊਖ਼͘͠ݕग़Ͱ͖ͳ͍
$PODMVTJPO w ϑϥϯεޠͰॻ͔ΕͨױऀͷΧϧςʹ͓͚Δ 0$3ޡΓͷݕग़ͱగਖ਼ํ๏ʹ͍ͭͯఏҊͨ͠ w ֶश࣌ʹϊΠζΛՃ͢Δ͜ͱͰ ೖྗจࣈྻͷϊΠζΛݕग़͠ɺగਖ਼Ͱ͖ΔΑ͏ʹͨ͠ w ૠೖޡΓͷݕʹ͓͍ͯྑͳ݁ՌͰͳ͔ͬͨ w
ࠓޙɺจࣈϨϕϧͷग़ྗͱ୯ޠϨϕϧͷใΛ Έ߹Θͤͯɺશମͷਫ਼Λ্͍ͤͨ͞ !16