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:Wronging a Right: Generating Better Errors...
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
Atsushi
November 30, 2018
130
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
文献紹介:Wronging a Right: Generating Better Errors to Improve Grammatical Error Detection
2018/11/30 文献紹介
長岡技術科学大学
自然言語処理研究室
Atsushi
November 30, 2018
More Decks by Atsushi
See All by Atsushi
文献紹介:Automated Evaluation of Out-of-Context Errors
atsumikan
0
110
文献紹介:Correction of OCR Word Segmentation Errors in Articles from the ACL Collection through Neural Machine Translation Methods
atsumikan
0
180
文献紹介:Auxiliary Objectives for Neural Error Detection Models
atsumikan
0
100
文献紹介:Low-resource OCR error detection and correction in French Clinical Texts
atsumikan
0
140
文献紹介:CMMC-BDRC Solution to the NLP-TEA-2018 Chinese Grammatical Error Diagnosis Task
atsumikan
0
140
文献紹介 : Fluency Boost Learning and Inference for Neural Grammatical Error Correction
atsumikan
0
200
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
170
文献紹介:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
atsumikan
0
150
文献紹介: Automatic Annotation and Evaluation of Error Types for Grammatical Error Correction
atsumikan
0
200
Featured
See All Featured
Unsuck your backbone
ammeep
672
58k
Introduction to Domain-Driven Design and Collaborative software design
baasie
1
900
Ruling the World: When Life Gets Gamed
codingconduct
0
280
Navigating Team Friction
lara
192
16k
Primal Persuasion: How to Engage the Brain for Learning That Lasts
tmiket
0
390
Claude Code のすすめ
schroneko
67
230k
The SEO identity crisis: Don't let AI make you average
varn
0
510
Become a Pro
speakerdeck
PRO
31
6k
Ethics towards AI in product and experience design
skipperchong
2
330
Measuring Dark Social's Impact On Conversion and Attribution
stephenakadiri
2
240
Large-scale JavaScript Application Architecture
addyosmani
515
110k
How to optimise 3,500 product descriptions for ecommerce in one day using ChatGPT
katarinadahlin
PRO
1
3.7k
Transcript
8SPOHJOHB3JHIU (FOFSBUJOH#FUUFS&SSPSTUP*NQSPWF (SBNNBUJDBM&SSPS%FUFDUJPO Ԭٕज़Պֶେֶࣗવݴޠॲཧݚڀࣨ ੁ३ࢤ จݙհ ݄ 4VEIBOTIV,BTFXB 1POUVT4UFOFUPSQ 4FCBTUJBO3JFEFM
1SPDFFEJOHTPGUIF$POGFSFODFPO&NQJSJDBM.FUIPETJO/BUVSBM -BOHVBHF1SPDFTTJOH QBHFTr#SVTTFMT #FMHJVN 0DUPCFS /PWFNCFS 1
"CTUSBDU w طଘͷBUUFOUJWFTFRTFRϞσϧΛ༻͠ ࣗવͳޡΓͷͷ܇࿅σʔλΛ࡞ w ඪ४తͳ#J-45.Ͱ4P5"ͷ݁ՌΛ্ճΔ w ਓखʹΑΔධՁͰࣗવͳޡΓͷจ͕ੜͰ͖͍ͯΔ͜ͱΛ ֬ೝͨ͠ !2
*OUSPEVDUJPO w ࣗಈతͳޡΓగਖ਼ͷੑೳɺ࣮༻ʹ͗͢Δ w ੑೳτϨʔχϯάσʔλʹΑ੍ͬͯݶ͞Ε͍ͯΔ w ߴ࣭ͳτϨʔχϯάσʔλΛௐୡ͢Δ͜ͱ͍͠ w طଘͷਓखʹΑΔΞϊςʔγϣϯσʔλΛੜͨ͠σʔλͰิ͏ w
ੜͨ͠σʔλͷޡΓͷ͕ݱ࣮ͷͱಉ͡߹ʹͷΈ༗ޮ w 3FJΒ 4.5Λར༻ͯ͠ޡΓͷͳ͍ςΩετʹޡΓΛੜ͠ɺ τϨʔχϯάηοτΛ࡞ͨ͠ !3
/FVSBMFSSPSHFOFSBUJPO w ޡΓΛ࡞͢ΔͨΊʹ༻͢ΔϞσϧ BUUFOUJWFTFRTFR// #BIEBOBVFUBM w %FDPEFS࣍ͷͭʹΑͬͯ୯ޠΛબ͢Δ w
BSHNBY ". w ࠷Մೳੑͷߴ͍୯ޠΛબ w 5FNQFSBUVSFTBNQMJOH 54 ɿ w BSHNBYͰબͰ͖ͳ͍رগޠΛબՄೳ w ֬ʹଟ༷ੑΛ࣋ͨͤΒΕΔ !4
/FVSBMFSSPSHFOFSBUJPO w Jɿ7PDBCVMBSZͷ୯ޠʹରԠ͢Δ֬ w НɿBSHNBYͱಉ͡ !5
/FVSBMFSSPSHFOFSBUJPO w 54ͰɺEFDPEFSͷ֤λΠϜεςοϓͰग़ྗ୯ޠΛબ w ੜ͞Εͨจ֬ͷ͍จͷՄೳੑ͕͋Δ w վળ͢ΔͨΊʹɺϏʔϜαʔν #4 Λ༻͍Δ w
ݱࡏͷλΠϜεςοϓ·Ͱʹੜ͞ΕͨOݸͷ࠷ྑͳ ༁ͷϦετΛอ࣋͢Δ w อ࣋ͨ݁͠Ռͷத͔Β#4Λ༻͍ͯ࠷ྑͷ͑ΛٻΊΔ !6
/FVSBMFSSPSHFOFSBUJPO w ޙॲཧ w ฤूڑΛ༻͍ͯϥϕϧ͚͢Δ w ਖ਼͍͠୯ޠʹz$z w ޡͬͨ୯ޠʹz*z w
ޡΓݕग़ثΛ܇࿅͢ΔͨΊͷσʔληοτʹ༻͍Δ w ॏෳͨ͠จͱͭҎ্ͷޡΓΛ࣋ͭจΛআ !7
&YQFSJNFOUT w ޡΓੜ෦ w 54 #4ͷ࣮ w Φʔϓϯιʔεͷ/.5ϑϨʔϜϫʔΫ #SJU[FUBM Λमਖ਼͠༻
w 54ɿН w ܇࿅σʔληοτ w QBSBMMFMDPSQVTWFSTJPOPG'$& :BOOBLPVEBLJTFUBM !8
&YQFSJNFOUT w ޡΓੜ෦ w ޡΓΛಋೖ͢Δίʔύε w '$&'JSTU$FSUJpDBUFPG&OHMJTI ,UPLFOT w
&71&OHMJTI7PDBCVMBSZ1SPpMF ,UPLFOT w 484JNQMF8JLJQFEJBͷαϒηοτ .UPLFOT !9
&YQFSJNFOUT w ޡΓݕग़෦ w ޡΓݕग़ʹ༻͢ΔϞσϧ w ඪ४తͳ#J-45. 4DIVTUFSBOE1BMJXBM
w TFRVFODFMBCFMMFSOFVSBMOFUXPSL 4- 3FJFUBM !10
&YQFSJNFOUT w ޡΓݕग़෦ w ܇࿅σʔληοτ w '$&ΞϊςʔγϣϯࡁΈτϨʔχϯάηοτ ఏҊख๏Ͱ૿ڧͨ͠τϨʔχϯάηοτ w
ධՁσʔληοτ w $P/--4IBSFEUBTL w $P/-- $P/-- w 'JSTU$FSUJpDBUFPG&OHMJTI '$& !11
3FTVMU !12
!13
!14
%JTDVTTJPO !15
!16
$PODMVTJPO w ݴޠֶशऀͷޡΓͷΛֶͼɺͦͷΑ͏ͳޡΓΛੜ͢ ΔͨΊʹ/.5Λ༻ͨ͠σʔλ૿ڧख๏Λఏࣔ w جຊతͳ#J-45.Ͱ͑͞طଘͷ4P5"ͷ݁ՌΛ্ճΓɺΑ Γྑ͍ϞσϧΛ༻͢Δ͜ͱʹΑͬͯ৽͍͠4P5"ͷ݁Ռ Λཱ֬ͨ͠ɻ w υϝΠϯ֎ͷσʔληοτͷޡΓΛར༻Ͱ͖Δ͜ͱΛ
࣮ূͨ͠ɻ !17