Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Sign up for free
Menu
Search
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Features
All features
Private URLs
Password Protection
Custom URLS
Scheduled publishing
Remove Branding
Restrict embedding
Deck Collections
Notes
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Explore
Featured decks
Featured speakers
Programming
Technology
Storyboards
Pricing
Search
Sign in
Sign up for free
[読み会] Knowledge distillation: A good teacher is...
Search
mei28
October 25, 2022
110
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
[読み会] Knowledge distillation: A good teacher is patient and consistent
読み会資料
Knowledge distillation: A good teacher is patient and consistent(CVPR2022)
mei28
October 25, 2022
More Decks by mei28
See All by mei28
[Human-AI Decision Making勉強会] 利用者の判断基準に合わせた説明は人の意思決定を誘導するか
mei28
0
64
[FAccT'26] Do User-Aligned Explanations Steer Human Decisions? Context-Dependent Influence and Ethical Implications
mei28
0
90
[Human-AI Decision Making勉強会] 説明を部分的に見せることで人に考えさせ、AIへの不適切な依存を減らす
mei28
0
160
[読み会] CHI2025論文紹介
mei28
1
120
[読み会] “Are You Really Sure?” Understanding the Effects of Human Self-Confidence Calibration in AI-Assisted Decision Making
mei28
0
190
[JSAI'24] 人間の判断根拠は文脈によって異なるのか?〜信頼されるXAIに向けた人間の判断根拠理解〜
mei28
2
960
[CHI'24] Fair Machine Guidance to Enhance Fair Decision Making in Biased People
mei28
0
150
[DEIM2024] 卓球の得点予測における重要要素の分析
mei28
0
83
[Human-AI Decision Making勉強会] 意思決定 with AIは個人vsグループで変わるの?
mei28
0
290
Featured
See All Featured
Joys of Absence: A Defence of Solitary Play
codingconduct
1
530
Lessons Learnt from Crawling 1000+ Websites
charlesmeaden
PRO
1
1.6k
Everyday Curiosity
cassininazir
0
340
The Invisible Side of Design
smashingmag
301
52k
Avoiding the “Bad Training, Faster” Trap in the Age of AI
tmiket
0
260
RailsConf & Balkan Ruby 2019: The Past, Present, and Future of Rails at GitHub
eileencodes
141
35k
Building Flexible Design Systems
yeseniaperezcruz
330
41k
ラッコキーワード サービス紹介資料
rakko
1
5.2M
How Fast Is Fast Enough? [PerfNow 2025]
tammyeverts
3
920
Designing for Timeless Needs
cassininazir
1
520
Agile Actions for Facilitating Distributed Teams - ADO2019
mkilby
0
300
Art, The Web, and Tiny UX
lynnandtonic
304
22k
Transcript
ಡΈձˏ ༶໌ ,OPXMFEHFEJTUJMMBUJPO "HPPEUFBDIFSJTQBUJFOUBOEDPOTJTUFOU
จใͱબཧ༝ • จใ • બཧ༝ • ࣝৠཹͷઃఆͷڭࢣͱੜెͷؔɼ ػցڭࣔͱྨࣅ͍ͯ͠ΔͨΊ
$7ٕज़ൃల͍ͯ͠Δ͚Ͳɼ࣮ࣾձͰ͍ͮΒ͍ • എܠɿେنϞσϧੑೳ͕ߴ͍͕ɼܭࢉίετɼઃඋίε τ͕ߴ͗͢Δ • ˠݱ࣮తʹখ͞ͳϞσϧͷํ͕͍উख͕ྑ͘ॏཁ • େنϞσϧͷೳྗΛ׆༻͢ΔͨΊʹɼੑೳҡ࣋ͨ͠··
ϞσϧΛܰྔԽ͢Δ͜ͱ͕ٻΊΒΕ͍ͯΔ
,%ʮڭࢣͱੜెͷϞσϧϚονϯάʯͱղऍ • ϞσϧͷܰྔԽʹͭͷख๏͕༗໊ • .PEFMQSVOJOH • ϞσϧͷҰ෦ΛΧοτ͢Δ͜ͱͰɼܰྔԽΛ͓͜ͳ͏ɽ • ϞσϧͷߏΛมߋͰ͖ͳ͍FH3FT/FUˠ.PCJMF/FU
• ,OPXMFEHFEJTUJMMBUJPO • ڭࢣʢେ͖Ίʣͱੜెʢখ͞ΊʣΛઃఆ͠ɽੜెڭࢣͱ ಉ͡ೳྗΛࢦ͢
ࣝৠཹɼೖྗσʔλͱֶश͕࣌ؒେࣄ • ͜͜ͰɼϞσϧΛܰྔԽ͢ΔͨΊʹɼࣝৠཹΛར༻ • ࣝৠཹʹ͓͍ͯԿ͕େࣄͳͷ͔Λݟ͚ͭΔ • ੜెͱڭࢣಉ͡ೖྗΛ༻͍Δ͜ͱ͕େࣄ • ೖྗσʔλσʔλ֦ு͢Δ΄Ͳྑ͍
• ֶश࣌ؒଟ͍ํ͕ྑ͍
ઃఆɾ࣮ݧઃఆ • ධՁࢦඪʹը૾ͷΫϥεྨਫ਼Λ༻͍Δ • େنϞσϧΛڭࢣͱઃఆ͠ɼྨਫ਼Λམͱͣ͞ʹɼϞσ ϧͷܰྔԽΛࢦ͢ɽ • ڭࢣɿ#J5ͷࣄલֶशϞσϧ •
ੜెɿ3FT/FU • σʔληοτछྨ খதنͭɼେنͭ • Ϋϥε σʔλαΠζ
ڭࢣͱੜెͷग़ྗΛ͚ۙͮΔ • ࣝৠཹͷֶशʹ༻͍ΔଛࣦͰ,-μΠόʔδΣϯε༻͍Δ • ͦΕͧΕڭࢣͷ༧ଌɼੜెͷ༧ଌ • Ϟσϧͷग़ྗࣗମɼԹ͖ιϑτϚοΫε pt ,
ps KL(pt ||ps ) = ∑ i∈C [−pt,i log ps,i + pt,i log pt,i ]
ϋΠύϥઃఆͱ͔࣮ݧͷখٕͨͪ • ࠷దԽख๏ͱͯ͠"EBNΛ༻͍ΔɽϋΠύϥσϑΥϧτ • εέδϡʔϥ$PTJOFEFDBZΛ༻͍Δ • ֶश҆ఆͷͨΊɼޯͷେ͖͞Λʹ੍ݶ • όοναΠζΛʹઃఆɽͨͩ͠*NBHF/FU
.JYVQʹΑͬͯσʔλ֦ுΛ͢Δ • ͭͷ܇࿅σʔλͷϖΞΛࠞ߹ͯ͠৽͘͠σʔλΛ࡞<> • ܇࿅σʔλ͕ϛοΫε͞ΕΔ͚ͩͰͳ͘ɼϥϕϧϛοΫε • • ͔ΒҰ༷ʹαϯϓϦϯά͢Δ
ຊՈ X = λX1 + (1 − λ)X2 , y = λy1 + (1 − λ)y2 λ ∈ [0,1] λ ∼ Beta(α, α) IUUQTBSYJWPSHBCT
ը૾ͷલॲཧ • σʔλͷલॲཧͰJODFQUJPOTUZMFDSPQΛ༻͍Δ • ݩσʔλͷը૾αΠζͷ ͰϥϯμϜʹΫϦοϓ • ΫϦοϓͨ͠ޙϥϯμϜͳΞεϖΫτൺʹม
• มͨ͋͠ͱݻఆαΠζ Y ʹϦαΠζ IUUQTBSYJWPSHBCT
ओுͷ͓͞Β͍ • ੜెͱڭࢣಉ͡ೖྗΛ༻͍Δ͜ͱ͕େࣄ • ೖྗσʔλσʔλ֦ு͢Δ΄Ͳྑ͍ • ֶश࣌ؒଟ͍ํ͕ྑ͍ • ϩόετੑΛݟΔͨΊʹখதنσʔληοτͭΛར༻
• ަབྷҼࢠΛഉআ͢ΔͨΊʹ͞·͟·ͳϋΠύϥͰ࣮ݧ 1BUJFOU UFBDIFS $POTJTUFOU UFBDIFS
$POTJTUFOUUFBDIFS Ұ؏ͨ͠ڭࢣ ʹ͍ͭͯΈΔ • ڭࢣɼੜెͷઃఆΛͭ༻ҙ͢Δɽ • 'JYFEUFBDIFSɿڭࢣͷग़ྗ͕ݻఆ͞Ε͍ͯΔ • *OEFQFOEFOUOPJTFɿڭࢣɼੜెͰҟͳΔೖྗ
• $POTJTUFOUUFBDIJOHɿڭࢣͱੜెͰಉ͡ೖྗ • 'VODUJPONBUDIJOHɿ$POTJTUFOUUFBDIJOHʴσʔλ֦ு • Լೋ͕ͭҰ؏ͨ͠ڭࢣઃఆΛද͍ͯ͠Δɽ
Ұ؏ͨ͠ڭࢣਫ਼͕ߴ͍ଞաֶशΛى͜͢
1BUJFOUUFBDIFS ڧ͍ڭࢣ ʹ͍ͭͯΈΔ • Ұൠతͳڭࢣ͋Γֶशͷ߹ɼϥϕϧʹର͠ɼը૾͕େ͖͘ ΉՄೳੑ͕͋Δɽ • ͜͜ͰɼࣝৠཹڭࢣͱੜెͷؔϚονϯάͱղऍ •
Ұ؏ͯ͠ಉ͡ೖྗΛ༩͑ΔͳΒɼೖྗࣗମΜͰ͍͍ ˠσʔλ֦ு͕༗ޮͳͷͰ • σʔλ֦ுΛߦ͍ɼେ͖ͳΤϙοΫͰֶश͢ΔͱΑΓੜె Ϟσϧͷੑೳ্͕͢Δ͜ͱΛݕূ͢Δ
աֶशΛճආ͠ͳ͕Βɼੑೳ্͕͢Δ • ͍ઢʢڭࢣʣͱಉͷੑೳʹ౸ୡ • ઢʢసҠֶशʣΑΓ࠷ऴతʹੑೳ͕ྑ͍ • ΦϨϯδʢ͔ΒֶशʣΑΓੑೳ͕͍͍ˠɹࣝৠཹͷޮՌ͕͋Δ
େنσʔληοτʢ*NBHF/FUʣͰࣝৠཹ • Ұ؏ͨ͠ڭࢣաֶश͍ͯ͠ͳ͍ • ·ͨσʔλ֦ு͢Δͱগͳ͍Τϙο ΫͰߴ͍ਫ਼Λୡ • ˠJUFSBUJPOಉ͡
ҟͳΔղ૾Ͱࣝৠཹ͕͏·͍͘͘ • ڭࢣ͕ߴ͍ղ૾ˠੜె͕͍ղ૾ͷํ͕ੑೳ͕ྑ͘ͳΔ
4IBNQPPΛֶͬͯशͨ͠Βߴʹͳͬͨ • σʔλ֦ுʹΑͬͯੑೳ্ͨ͠ ͕ɼͦͷֶश͕࣌ؒ৳ͼͯ͠·͏ • "EBN͔ΒΑΓڧྗͳ࠷దԽख๏Ͱ ͋Δ4IBNQPPΛͬͯΈΔɽ • ˠֶश͕ഒʹͳͬͨʂ
IUUQTBSYJWPSHBCT
సҠֶश͔ͭͬͯࣝৠཹ͍͍Μ͡Όͳ͍ • 1BUJFOUUFBDIFSͷͱ͖ɼసҠֶ श༗ޮ͔ͩͬͨΒࣝৠཹ స Ҡֶश͍͍ͷͰ • ֶशॳظసҠֶशΑ͔͚ͬͨ
Ͳɼ࠷ऴతʹٯస
ࣝৠཹͬͺΓ༗ޮ • ࣝৠཹͬͯແ͍͍ͯͬͯ͘͜ ͱͳ͍ʁ • ࣝৠཹ͕ͳ͍ͱաֶशΛى͜͢
·ͱΊͱײ • ࣝৠཹͰॏཁͳཁૉͱͯ͠ɼ • ɽڭࢣͱੜెͷೖྗ͕ಉ͡ • ɽσʔλ֦ுΛͨ͘͞Μ͢Δ • ɽֶशΤϙοΫΛ૿͢ɹɹɹɹ͜ͱʹ͍ͭͯࣔͨ͠ɽ
• ࣝৠཹͷ৽ख๏ΛఏҊͨ͠Θ͚Ͱͳ͍͕ɼطଘͷϞσϧΛ༻ ͍ͯɼܰྔϞσϧ͕4P5"ΛऔΕΔ͔͠Εͳ͍ϩϚϯΛײ͡Εͨ