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:Automated Evaluation of Out-of-Context Errors
Search
Atsushi
February 06, 2019
Technology
0
99
文献紹介:Automated Evaluation of Out-of-Context Errors
2019/02/06 文献紹介
長岡技術科学大学
自然言語処理研究室
Atsushi
February 06, 2019
Tweet
Share
More Decks by Atsushi
See All by Atsushi
文献紹介:Correction of OCR Word Segmentation Errors in Articles from the ACL Collection through Neural Machine Translation Methods
atsumikan
0
170
文献紹介:Auxiliary Objectives for Neural Error Detection Models
atsumikan
0
94
文献紹介:Wronging a Right: Generating Better Errors to Improve Grammatical Error Detection
atsumikan
0
130
文献紹介:Low-resource OCR error detection and correction in French Clinical Texts
atsumikan
0
130
文献紹介:CMMC-BDRC Solution to the NLP-TEA-2018 Chinese Grammatical Error Diagnosis Task
atsumikan
0
130
文献紹介 : Fluency Boost Learning and Inference for Neural Grammatical Error Correction
atsumikan
0
190
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
160
文献紹介:The Effect of Error Rate in Artificially Generated Data for Automatic Preposition and Determiner Correction
atsumikan
0
140
文献紹介: Automatic Annotation and Evaluation of Error Types for Grammatical Error Correction
atsumikan
0
180
Other Decks in Technology
See All in Technology
マルチロールEMが実践する「組織のレジリエンス」を高めるための組織構造と人材配置戦略
coconala_engineer
2
400
Oracle Cloud Infrastructure:2026年2月度サービス・アップデート
oracle4engineer
PRO
0
220
男(監査)はつらいよ - Policy as CodeからAIエージェントへ
ken5scal
5
730
オンプレとGoogle Cloudを安全に繋ぐための、セキュア通信の勘所
waiwai2111
3
1.1k
技術的負債の泥沼から組織を救う3つの転換点
nwiizo
7
2.1k
Datadog Cloud Cost Management で実現するFinOps
taiponrock
PRO
0
140
型を書かないRuby開発への挑戦
riseshia
0
180
マネージャー版 "提案のレベル" を上げる
konifar
17
12k
「データとの対話」の現在地と未来
kobakou
0
1.3k
「使いにくい」も「運用疲れ」も卒業する UIデザイナーとエンジニアが創る持続可能な内製開発
nrinetcom
PRO
1
780
Claude Codeの進化と各機能の活かし方
oikon48
7
2k
【SLO】"多様な期待値" と向き合ってみた
z63d
2
300
Featured
See All Featured
Side Projects
sachag
455
43k
Practical Tips for Bootstrapping Information Extraction Pipelines
honnibal
25
1.8k
The B2B funnel & how to create a winning content strategy
katarinadahlin
PRO
1
290
The Art of Delivering Value - GDevCon NA Keynote
reverentgeek
16
1.9k
Into the Great Unknown - MozCon
thekraken
40
2.3k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
12
1.4k
Sam Torres - BigQuery for SEOs
techseoconnect
PRO
0
210
The Hidden Cost of Media on the Web [PixelPalooza 2025]
tammyeverts
2
240
What the history of the web can teach us about the future of AI
inesmontani
PRO
1
450
My Coaching Mixtape
mlcsv
0
63
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
35
3.4k
How to Build an AI Search Optimization Roadmap - Criteria and Steps to Take #SEOIRL
aleyda
1
1.9k
Transcript
"VUPNBUFE&WBMVBUJPOPG 0VUPG$POUFYU&SSPST Ԭٕज़Պֶେֶࣗવݴޠॲཧݚڀࣨ ੁ३ࢤ จݙհ ݄ 1BUSJDL)VCFS +BO/JFIVFT "MFY8BJCFM 1SPDFFEJOHTPGUIF&MFWFOUI*OUFSOBUJPOBM$POGFSFODFPO-BOHVBHF
3FTPVSDFTBOE&WBMVBUJPO -3&$ QBHFTr .JZB[BLJ +BQBO .BZ
"CTUSBDU w ςΩετཧղλεΫͷͨΊͷ ܭࢉϞσϧΛධՁ͢Δ৽͍͠ΞϓϩʔνΛఏҊ w ࣗಈతͳվมख๏ʹΑΓେنͳσʔληοτ Λ༻Մೳʹ w 5&%ίʔύεΛվม͠ධՁ
*OUSPEVDUJPO w ࣗಈԻೝࣝ "43 χϡʔϥϧػց༁ /.5 γεςϜͷଟ͘Ͱ Ұʹจͣͭॲཧ w จ͚ͩͰจ຺֎ͷޡΓͷࣝผ͕ࠔ
w ྫ "43γεςϜͰੈքܦࡁʹؔ͢ΔจΛॻ͖ى͜͢߹ w ਖ਼ 5IFSFBSFHSFBUPQQPSUVOJUJFTJOUIFGBSFBTU w ޡ 5IFSFBSFHSFBUPQQPSUVOJUJFTJOUIFGPSFTU w
Իతʹࣅ͍ͯΔͨΊ GBSFBTU͕GPSFTUͱޡೝࣝ w จ຺্ GPSFTUਖ਼͘͠ͳ͍୯ޠ *OUSPEVDUJPO
*OUSPEVDUJPO w ৽͍͠ςΩετཧղλεΫΛઃܭ w จ຺֎ͷޡΓΛࣝผ͢ΔλεΫ w ࣗಈԽ͞Εͨσʔληοτ࡞खॱΛఏҊ w จ຺֎ͷޡΓΛࣝผ͢ΔϞσϧΛධՁ͢ΔͨΊ ͷσʔληοτ
5BTL w จ຺֎ͷޡΓΛࣝผ͢ΔλεΫ w ֤୯ޠ͕ޡΓͰ͋Δ͔൱͔ͷ όΠφϦܥྻϥϕϦϯά w ޡΓҐஔσʔληοτશମʹࢄ͞Ε͍ͯΔ ͨΊݻఆҐஔΞϓϩʔνΑΓλεΫ͕ෳࡶ
5BTL w 8F IBWF JO
IJHIFS FEVDBUJPO B USJMMJPO EPMMBST PG TUVEFOU EFCU <ʜ> 8F IBWF B MPU PG TIPQQJOH 0VS FDPOPNZ HSFX <ʜ> PO UIF CBDL PG DPOTVNFST NBTTJWFMZ PWFS CPSSPXJOH
%BUB.PEJpDBUJPO w ϞσϧΛධՁ͢ΔͨΊͷσʔληοτͷ࡞खॱ w طଘͷσʔληοτʹจ຺֎ͷޡΓΛૠೖ w ಛʹڧྗͳจ຺্ͷಛΛ࣋ͭจষΛ༻͍Δ w શʹࣗಈԽ͞Ε͓ͯΓ ਓखΛඞཁͱ͠ͳ͍
w େنίʔύεͷվมՄೳ
%BUB.PEJpDBUJPO w σʔληοτͷ࡞खॱ %BUBTFU'JMUFSJOH 1BSUPG4QFFDI5BHHJOH $BOEJEBUF4FMFDUJPO
"QQFBSBODF8JOEPX
%BUBTFU'JMUFSJOH w σʔληοτͷ࣭Λ্ͤ͞ΔͨΊʹ ޠະຬͷ୯ޠͰߏ͞ΕΔจষΛআڈ͢Δ
1BSUPG4QFFDI5BHHJOH w ஔ͞ΕΔ୯ޠͷબͷͨΊʹ ࢺΛ༩͢Δ w จ຺ಛʹ໊ࢺʹͱͬͯॏཁ1BQFSOPFUBM w ໊ࢺͷஔʹযΛͯΔ
w จষதͷ໊ࢺͷϦετΛऔಘ͢Δ
$BOEJEBUF4FMFDUJPO w ஔ͞ΕΔ୯ޠΛબ͢Δ w จষͷҰ෦ʹෳճग़ݱ͢Δ໊ࢺ͕จ຺Λܾఆ w ໊ࢺϦετͷதͰ ྡ͢Δ୯ޠҎʹෳճग़ݱ͠ ໊ͨࢺͷ͏ͪ ࠷ޙͷ໊ࢺͷҐஔΛஔҐஔͷީิͱ͢Δ
w ྫʣ໊ࢺ" ໊ࢺ# ໊ࢺ$ ໊ࢺ% ໊ࢺ& ໊ࢺ' ໊ࢺ( ໊ࢺ$ ໊ࢺ* ໊ࢺ$ w ஔҐஔ ީิ͔ΒϥϯμϜʹબ͢Δ
"QQFBSBODF8JOEPX w ஔ͢Δ୯ޠΛબ͢Δ w ݩͷσʔληοτ্Ͱ ग़ݱճ͕ಉఔͷ୯ޠΛબ w ͜ͷख๏ʹΑΓ ҎԼͷಛΛಘΔ
Ұൠతͳ୯ޠ͕සͳ୯ޠʹஔ͞ΕΔ͜ͱΛճආ ୯ޠͷग़ݱճҰൠతʹҙຯͱ͕ؔͳ͍ͨΊ ஔ͞Εͨ୯ޠͱݩͷ୯ޠͷؒͰؔΛ࣋ͨͳ͍ w ͞Βʹ੍͓࣌Αͼʹؔ͢ΔϑΟϧλϦϯάΛߦ͏
/PXOPUJDF JOBQFSJPEXIJDIJTEPNJOBUFECZB NPOPQPMBSXPSME ZPVIBWFpYFEBMMJBODFTr/"50 UIF8BSTBX1BDU"pYFEQPMBSJUZPGQPXFSNFBOT pYFEBMMJBODFT#VUBNVMUJQMFQPMBSJUZPGQPXFS NFBOTTIJGUJOHBOEDIBOHJOHBMMJBODFT"OEUIBU`TUIF XPSMEXF`SFDPNJOHJOUP JOXIJDIXFXJMMJODSFBTJOHMZ
TFFUIBUPVSNBSLFUFSTBSFOPUpYFE .PEJpDBUJPOEBUBFYBNQMF ਖ਼ʣBMMJBODFT ޡʣNBSLFUFST
&WBMVBUJPO w σʔληοτ w &OHMJTI5&%5BMLDPSQVT SFMFBTFEGPSUIF*84-5 $FUUPMPFUBM w
ϕʔεϥΠϯ w -BOHVBHF.PEFM w #JOBSZ$MBTTJpDBUJPO.PEFM w )VNBO#BTFMJOF
#BTFMJOF w -BOHVBHF.PEFM w 3//Λ༻͍ͨڭࢣͳֶ͠शϞσϧ w #JOBSZ$MBTTJpDBUJPO.PEFM w OFVSBM-.ͷग़ྗΛͱͨ͠ڭࢣ͋ΓֶशϞσϧ w
)VNBO#BTFMJOF w จΛݟͯจ຺ʹ߹Θͳ͍୯ޠΛਓ͕ؒݟ͚ͭΔ w ςετηοτ͔ΒϥϯμϜʹબΜͩจʹର͠ ਓͷࢀՃऀ͕ճ
None
&WBMVBUJPO
$PODMVTJPO w จ຺Λҙࣝͨ͠ܭࢉϞσϧΛධՁ͢ΔͨΊͷ ৽͍͠λεΫΛఏࣔ w จ຺ͷҙຯදݱʹযΛ߹Θͤͨจ຺ೝࣝϞσϧͷͨΊ ͷแׅతͳϕϯνϚʔΫΛ࡞ w ہॴతͳτʔΫϯͷΈΛ༻͢ΔϞσϧͰ ஔ͞ΕͨτʔΫϯͷਪఆ͕ࠔͰ͋Δ͜ͱ͕ࣔ͞Εͨ
w λεΫͷύϑΥʔϚϯεΛ্ͤ͞ΔͨΊʹ ঢ়گʹԠͨ͡ΞϓϩʔνͷఏҊ͕ඞཁͰ͋Δ