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
Kaggle Google Quest Q&A Labeling - 23th place s...
Search
Shuhei Goda
February 28, 2020
Technology
4.2k
4
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Kaggle Google Quest Q&A Labeling - 23th place solution
Shuhei Goda
February 28, 2020
More Decks by Shuhei Goda
See All by Shuhei Goda
Turing × atmaCup #18 - 1st Place Solution
hakubishin3
0
1.3k
ジョブマッチングサービスにおける相互推薦システムの応用事例と課題
hakubishin3
3
1.2k
とある事業会社にとっての Kaggler の魅力
hakubishin3
9
3.3k
課題の解像度が荒かったことで意図した改善ができなかった話
hakubishin3
3
1.2k
Wantedly におけるマッチング体験を最大化させるための推薦システム
hakubishin3
4
1.4k
Recommendation Industry Talks #1 Opening
hakubishin3
1
480
会社訪問アプリ「Wantedly Visit」での シゴトに関する興味選択機能と推薦改善
hakubishin3
0
790
論文紹介: Improving Implicit Feedback-Based Recommendation through Multi-Behavior Alignment(Xin Xin et al., 2023)
hakubishin3
0
740
Feedback Prize - English Language Learning における擬似ラベルの品質向上の取り組み
hakubishin3
1
1.2k
Other Decks in Technology
See All in Technology
#jawssonic2026 あの時代が悪かった ~動かなかったSageMakerと共に迎えたイベント当日~
ktkn1129
0
130
PfEingのアプローチで働こう
rindrics
0
180
AIとペアプロを始める。人とのペアプロをやめる。ペアプロの良さを改めて知る。もっと好きになった。 / Rediscovering Pair Programming
honyanya
1
250
2026-09-04 SRE Tech Talk #15 怠惰なTerraform / Lazy Terraform
masasuzu
0
200
Microsoft 365 Copilot chat -tekoälypalvelun tietosuojaongelmat
hponka
0
430
Kiro Crewしか勝たん!?
miu_crescent
PRO
0
180
スクラムで身についていた動き方を、XPで捉え直してみた
codmoninc
PRO
1
250
AI時代のAPI品質を支えるガードレール / API Guardrails for API quality in the AI era
yokawasa
1
120
AIエージェントの開発・提供におけるセキュリティリスクの論点と対策
flatt_security
1
580
AndroidでHDRメディアを「壊さずに」扱う
chigichan24
0
410
深夜のクラウド懺悔室 1:29:300 or 1:0:0
kazzpapa3
0
180
その指示、Bobにしっかり伝わってる?初心者でも即実践できる プロンプト/コンテキスト設計のコツ
muehara
1
170
Featured
See All Featured
svc-hook: hooking system calls on ARM64 by binary rewriting
retrage
2
560
How Software Deployment tools have changed in the past 20 years
geshan
1
34k
The AI Search Optimization Roadmap by Aleyda Solis
aleyda
1
6.1k
Accessibility Awareness
sabderemane
1
200
Thoughts on Productivity
jonyablonski
76
5.4k
How to Ace a Technical Interview
jacobian
281
24k
The Curious Case for Waylosing
cassininazir
1
500
The Limits of Empathy - UXLibs8
cassininazir
1
640
RailsConf 2023
tenderlove
30
1.5k
Stop Working from a Prison Cell
hatefulcrawdad
274
21k
Bootstrapping a Software Product
garrettdimon
PRO
306
120k
SEOcharity - Dark patterns in SEO and UX: How to avoid them and build a more ethical web
sarafernandez
0
270
Transcript
©2020 Wantedly, Inc. 23th place solution Kaggle Google Quest Q&A
Labeling লձ Feb 28, 2020 - Shuhei Goda - @jy_msc
©2020 Wantedly, Inc. Team - The Hand Shuhei Goda @jy_msc
Visit Engineering Team at Wantedly Naomichi Agata @agatan_ People Engineering Team at Wantedly
©2020 Wantedly, Inc. Model Pipeline #FSUCBTF VODBTFE -JHIU(#. #FSUCBTF VODBTFE
Settings ɾ3fold with GroupKFold ɾBCE + margin ranking loss ɾ3epoch Settings ɾmax_depth=1 ɾlr=0.1 Meta features ɾtext length ɾstackexchange Text data ɾquestion_title ɾquestion_body ɾanswer 1SF1SPDFTT 2BOE" 1SF1SPDFTT POMZ2 ɾquestion_title ɾquestion_body ɾquestion_title ɾquestion_body ɾanswer Settings ɾhtml escape ɾhead+tail truncation
©2020 Wantedly, Inc. ɾHTMLจࣈྻͷΞϯΤεέʔϓ Pre-Process IUUQTXXXLBHHMFDPNDHPPHMFRVFTUDIBMMFOHFEJTDVTTJPO
©2020 Wantedly, Inc. ɾςΩετσʔλͷ݁߹ͱτϦϛϯά ɹɾ[CLS] + question_title + [SEP] +
question_body + [SEP] + answer ɾquestion_body ͱ answer ͕ࢦఆͷ͞Λ͑ͨ߹, ͔྆ΒಉαΠζΛτϦϛϯά Pre-Process IUUQTBSYJWPSHBCT
©2020 Wantedly, Inc. ɾBert-base (uncased) ɹɾޙΖ4ͭͷӅΕͷग़ྗΛ༻ https://arxiv.org/abs/1905.05583 ɹɾQAؒͷSEP tokenͷग़ྗΛ༻ Model
Architecture
©2020 Wantedly, Inc. ɾLabel weight ɹɾ؆୯ͦ͏ͳλεΫweightΛখ͘͞, ෆۉߧͰͦ͠͏ͳλεΫweightΛେ͖͘ ɹɾgpyoptͰweightͷ୳ࡧΛࢼͨ͠Έ͕ͨ, Լهͷ୯७ͳΓํ͕࠷ྑ͔ͬͨ Loss
function Label weight ͋Γ Public: 0.45979, Private: 0.41440 Label weight ͳ͠ Public: 0.43455, Private: 0.40602
©2020 Wantedly, Inc. ɾBCE + margin ranking loss (1 :
1) ɹɾϛχόονΛ2ͭʹׂͯ͠ margin ranking loss Λܭࢉ Loss function BCE + margin ranking loss Public: 0.45979, Private: 0.41440 BCE Public: 0.44006, Private: 0.40668
©2020 Wantedly, Inc. ɾQuestion Model ɹɾQ༻ͷλεΫΛQuestion text͚ͩΛͬͯղ͘ ɹɾΠϯϓοτQ͚ͩͰ͍͍ͷͰ, Qͷtruncationͷྔ͕ݮΔ (Qͷใྔ͕૿͑Δ)
Training Q model + Q and A model Public: 0.45979, Private: 0.41440 Q and A model × 2 (seed average) Public: 0.44298, Private: 0.40613
©2020 Wantedly, Inc. ɾLightGBM ɹɾmax_depth=1, lr=0.1 ɹɾmeta features ɹɹɾtext length
(question, answer) ɹɹɾmeta data from stackexchange (Score, View, FavoriteCount, …) Post-Process LightGBM Public: 0.45979, Private: 0.41440 Simple binning without meta features Public: 0.45282, Private: 0.41387
©2020 Wantedly, Inc. Why we used LightGBM? 1. Simple binning
method ɹɾ༧ଌΛࢄԽ͢Δ͜ͱͰ Spearman’s correlation ͕ྑ͘ͳΔ͜ͱʹؾͮ͘ ɹɾtarget͝ͱʹϏϯαΠζΛࣄલʹઃఆͯ͠Ϗϯೋϯά ɹɾϏϯαΠζݻఆʹ্ͨ͠ͰBertͷ֤epochͷग़ྗΛweighted average (weight࠷దԽ)
©2020 Wantedly, Inc. Why we used LightGBM? 2. Optimize bin-size
and weights ɹɾϏϯαΠζ࠷దͳΛ͍ͨ͘ͳͬͨ ɹɾϏϯαΠζͱweightsͷಉ࣌࠷దԽ্͕ͨ͠ख͍͔͘ͳ͍ ɹɾ࠷దͳϏϯαΠζ༧ଌͷܗʹΑܾͬͯ·Δ. ֤foldͷ࠷దͳϏϯαΠζͷฏۉͱ weighted averageޙͷ༧ଌ࠷దͳͷ͔Βဃ͢Δ
©2020 Wantedly, Inc. Why we used LightGBM? 3. LightGBM ɹɾϏϯαΠζͱweightsͷಉ࣌࠷దԽ͍ͨ͠
ɹɾmeta features͍͍ͨ ɹɾGBDTσʔλΛׂׂͯ͠ޙͷྖҬʹ࠷దͳΛׂΓͯΔख๏ ɹɹˠ ઙ͍߹Ϗϯχϯάͱಉ༷ͷࢄԽ͕Ͱ͖ΔΜ͡Όͳ͍͔ max_depth=2 max_depth=8
©2020 Wantedly, Inc. 4. LightGBM (parameter tuning) ɹɾࢄԽ͢Δ΄Ͳscore͕ྑ͘ͳΔͷͰ, ߏΛۃྗγϯϓϧʹ͍ͨ͠ ɹɾtrainσʔλΛׂͯ͠࠷దͳύϥϝʔλΛݟ͚ͭΔ
ɹɾmax_depthΛҰ൪খ͘͞, lrΛۃྗେ͖ͨ͘͠ํ͕score͕ྑ͘ͳͬͨ Why we used LightGBM?
©2020 Wantedly, Inc. ɾsample weightͷઃఆ ɾhostͷ୯ޠΛΠϯϓοτͷઌ಄ྻʹஔ͘ ɾnew tokenͷՃ ɾBert-base casedΛ͏
ɾtexͷίʔυϒϩοΫΛྗٕͰফڈ Didn’t work for us
©2020 Wantedly, Inc. Discussion: https://www.kaggle.com/c/google-quest-challenge/discussion/129904#742302 Kernel: https://www.kaggle.com/shuheigoda/23th-place-solusion Links
©2020 Wantedly, Inc. https://www.wantedly.com/projects/375150 We are hiring !