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
Variational Auto Encoderでの Disentangled表現
Search
Kosuke Miyoshi
November 03, 2016
Research
640
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Variational Auto Encoderでの Disentangled表現
※ Speaker Deck内ではURLリンクがクリックできないので、クリックしたい場合はpdfをダウンロードしてください
Kosuke Miyoshi
November 03, 2016
More Decks by Kosuke Miyoshi
See All by Kosuke Miyoshi
Representation Learning with Contrastive Predictive Coding
miyosuda
1
240
Sutton "Reinforcement Learning" 2nd Edition Ch13: Policy Gradient Methods
miyosuda
0
220
Sutton "Reinforcement Learning" 2nd Edition Ch7: n-step Bootstrapping
miyosuda
0
120
Sutton "Reinforcement Learning" 2nd Edition Ch6: TD-learning
miyosuda
0
120
SCAN
miyosuda
0
890
TensorFlow & DeepMind Lab & UNREAL
miyosuda
1
2.6k
Other Decks in Research
See All in Research
LLM の Attention 機構まとめ — 数式・計算量・メモリ
puwaer
8
2.2k
第66回コンピュータビジョン勉強会@関東 Epona: Autoregressive Diffusion World Model for Autonomous Driving
kentosasaki
0
640
オーストリア流 都市の公共交通サービス水準評価@公共交通オープンデータ最前線2026
trafficbrain
0
190
NII S. Koyama's Lab Research Overview AY2026
skoyamalab
0
340
コーディングエージェントとABNを再考
hf149
2
730
「AIとWhyを深堀る」をAIと深堀る
iflection
0
500
NLP colloquium: AI Safety Survey
kanekomasahiro
0
750
Apache Gravitinoで実現する Icebergカタログ統合とアクセスの一元化
matsumooon
0
300
Can We Teach Logical Reasoning to LLMs? – An Approach Using Synthetic Corpora (AAAI 2026 bridge keynote)
morishtr
1
260
Fukui Shibiten 39 - AI Art
butchi
0
130
事後確率分布の共分散について
koide3
0
150
AI Agentの精度改善に見るML開発との共通点 / commonalities in accuracy improvements in agentic era
shimacos
6
1.7k
Featured
See All Featured
Making the Leap to Tech Lead
cromwellryan
135
9.9k
Bootstrapping a Software Product
garrettdimon
PRO
307
120k
Why Your Marketing Sucks and What You Can Do About It - Sophie Logan
marketingsoph
0
170
Accessibility Awareness
sabderemane
1
140
How to build an LLM SEO readiness audit: a practical framework
nmsamuel
1
790
SEO for Brand Visibility & Recognition
aleyda
0
4.6k
How Fast Is Fast Enough? [PerfNow 2025]
tammyeverts
3
620
Un-Boring Meetings
codingconduct
0
320
The World Runs on Bad Software
bkeepers
PRO
72
12k
For a Future-Friendly Web
brad_frost
183
10k
Crafting Experiences
bethany
1
190
Game over? The fight for quality and originality in the time of robots
wayneb77
1
200
Transcript
7BSJBUJPOBM"VUP&ODPEFSͰͷ %JTFOUBOHMFEදݱ OBSSBUJWFOJHIUTגࣜձࣾࡾ߁༞ શΞʔΩςΫνϟୈճ4*(8#"Φϑձ
&BSMZ7JTVBM$PODFQU-FBSOJOHXJUI 6OTVQFSWJTFE%FFQ-FBSOJOH Irina Higgins, Loic Matthey, Xavier Glorot, Arka Pal,
Benigno Uria, Charles Blundell, Shakir Mohamed, Alexander Lerchner Google DeepMind, 2016 ͱΓ͋͛Δจ ˞όʔδϣϯʹΑͬͯύϥϝʔλ͕େ͖͘ҧ͍ͬͯΔɻ࠷৽ͷWΛར༻
%JTFOUBOHMFE SFQSFTFOUBUJPO w ҙຯۭؒͷදݱΛɺҐஔɺαΠζɺFUDʜͳͲͷߏ ཁૉͷύϥϝʔλʹ͔Εͨঢ়ଶͰ֫ಘ͍ͨ͠ w 7BSJBUJPOBM"VUP&ODPEFSͷજࡏมͰ͜ΕΒͷ දݱΛڭࢣͳ͠Ͱ֫ಘͰ͖ͳ͍͔ʁ
7BSJBUJPOBM"VUP&ODPEFS 9 [ 9` FODPEFS EFDPEFS αϯϓϦϯά
7BSJBUJPOBM"VUP&ODPEFS zͷࣄޙۙࣅq(z|x) (=Τϯίʔμͷग़ྗ ͱ N(0,I)ͱͷڑΛ࠷খԽ ෮ݩޡࠩΛ࠷খԽ
Ќͩͱzͷࣄޙۙࣅq(z|x)ΛΑΓ N(0,I)ʹ͚ۙͮΑ͏ͱ͢Δ ਖ਼ଇԽ߲ʹЌΛಋೖ Ќͷࢹ֮࿏ʹ͋ΔύϥϝʔλʁΛۙࣅ
w Y࠲ඪ Z࠲ඪͦΕͧΕύλʔϯ w TDBMFύλʔϯ w ճసύλʔϯ w ΦϒδΣΫτͷܗ ପԁɺϋʔτɺ࢛֯
ύλʔϯ Yͷը૾ݸΛજࡏมݸͷ7"&Ͱֶश [ FODPEFS EFDPEFS αϯϓϦϯά ͷμ,σ
Ќ Ќ
N(0, I)ʹٵ͍ࠐ·Εͯ͠·͍ͬͯΔ જࡏมݸͷ͏ͪͷͻͱͭͷЖ͚ͩΛม͑ͯΓΛݻఆ ֤જࡏม͕ύϥϝʔλ͝ͱʹ ͔Ε͍ͯΔ જࡏมͷࢄ જࡏมͷЖΛdͰ ಈ͔͢
%FFQ.JOEʹΑΔͦͷଞͷ࣮ݧ݁Ռ IUUQUJOZVSMDPNKHCZ[LF "UBSJ %໎࿏FUD…
7"&Ͱ%JTFOUBOHMF͞ΕΔʹ w ม͕࿈ଓͰຶʹαϯϓϦϯά͞Ε͍ͯΔ w ཁૉ͕౷ܭతʹಠཱ w ੑ͕ݮΒ͞Ε͍ͯΔ
w ີʹαϯϓϦϯά͞Ε͍ͯΕɺೋ͕ͭผͷΦϒδΣΫτͩͱೝࣝͰ͖Δ
ಈ͖͕࿈ଓͰີʹαϯϓϦϯά͞Ε͍ͯΕɺ ಉ͡ΦϒδΣΫτͩͱೝࣝͰ͖Δ ͜ΕΒೋͭͷύυϧಉ͡ΦϒδΣΫτͩΖ͏͔ʁ
࠶ݱݕূ ETIBQF
࠶ݱݕূ #SFBLPVU ΫϦοΫͰ࠶ੜ wϒϩοΫͷมԽ͕ͭͬͨજࡏมʹ·͕ͨͬͯऔΕ͍ͯΔ จͷ༷ͳύυϧͷಈ͖είΞͷมԽ·ͩऔΕ͍ͯͳ͍ wЌΛڧ͘͢Δͱͬͯ͘Δજࡏมݮͬͯ͘Δ͕ɺಉ࣌ʹϘέ ۩߹͕େ͖͘ͳͬͯύυϧ͕ফ͑ͯ͠·͏ [ [
[