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
入門AlphaGo
Search
na-o-ys
April 22, 2016
Technology
5
3.8k
入門AlphaGo
"Mastering the game of Go with deep neural networks and tree search" の概要
na-o-ys
April 22, 2016
Tweet
Share
More Decks by na-o-ys
See All by na-o-ys
IoTと監視
naoys
1
780
RubyとJIT
naoys
0
160
将棋盤を画像認識したかった
naoys
0
1.5k
Rust で乗り換え案内
naoys
0
620
疎行列と Jaccard 類似度の高速計算
naoys
1
610
有理数集合の濃度
naoys
2
120
YARVの最適化について調べた
naoys
0
130
転職会議サービスのAWS移行記録
naoys
0
64
Anonymous Recursion in C++
naoys
0
420
Other Decks in Technology
See All in Technology
ObsidianをLLM時代のナレッジベースに! クリッピング→Markdown→CLI連携の実践
srvhat09
7
7.7k
Microsoft Fabric ガバナンス設計の一歩目を考える
ryomaru0825
1
160
第64回コンピュータビジョン勉強会「The PanAf-FGBG Dataset: Understanding the Impact of Backgrounds in Wildlife Behaviour Recognition」
x_ttyszk
0
260
Data Engineering Study#30 LT資料
tetsuroito
1
470
安定した基盤システムのためのライブラリ選定
kakehashi
PRO
3
160
Shadow DOM & Security - Exploring the boundary between light and shadow
masatokinugawa
0
590
AIでテストプロセス自動化に挑戦する
sakatakazunori
1
610
デザインレビューをAIにしてもらいたい!
tomuro
2
120
An introduction to Claude Code SDK
choplin
3
3k
Bill One 開発エンジニア 紹介資料
sansan33
PRO
4
13k
Maintainer Meetupで「生の声」を聞く ~講演だけじゃないKubeCon
logica0419
0
140
名刺メーカーDevグループ 紹介資料
sansan33
PRO
0
820
Featured
See All Featured
Practical Orchestrator
shlominoach
189
11k
Designing for Performance
lara
610
69k
BBQ
matthewcrist
89
9.7k
Building a Scalable Design System with Sketch
lauravandoore
462
33k
The Cost Of JavaScript in 2023
addyosmani
51
8.6k
I Don’t Have Time: Getting Over the Fear to Launch Your Podcast
jcasabona
32
2.4k
Learning to Love Humans: Emotional Interface Design
aarron
273
40k
Evolution of real-time – Irina Nazarova, EuRuKo, 2024
irinanazarova
8
840
[RailsConf 2023 Opening Keynote] The Magic of Rails
eileencodes
29
9.6k
Unsuck your backbone
ammeep
671
58k
個人開発の失敗を避けるイケてる考え方 / tips for indie hackers
panda_program
108
19k
Performance Is Good for Brains [We Love Speed 2024]
tammyeverts
10
980
Transcript
ೖAlphaGo 0x64ޠ ୈ07 “AI” @na_o_ys
͝ҙ • จʹॻ͔Ε͍ͯͳ͍ಠࣗௐࠪਪଌؚ͕· Ε·͢ • Ұߟͩͱࢥͬͯݟ͍ͯͩ͘͞
AlphaGoͱ • ॳΊͯϓϩع࢜ΛഁͬͨғޟAI
ୈҰ෦: AlphaGoʹࢸΔ·Ͱ
શใήʔϜ • ΦηϩɺνΣεɺকعɺғޟɺetc • ϥϯμϜੑ͕ແ͘ɺ࠷ળख͕ଘࡏ͢Δ • (ݪཧతʹ) ઌखඞউɾޙखඞউɾҾ͖͚
ήʔϜ • શ୳ࡧͰ࠷ળख͕ٻ·Δ • ܭࢉྔతʹෆՄೳ … ݱہ໘ 1खޙ 2खޙ
ධՁؔ • ൫໘Λ༩͑ΔͱείΞ (༧উͳͲ) Λฦؔ͢ • কعνΣεͳΒɺۨͷଛಘޮ͖ͷΛݩʹܭࢉ • ήʔϜͷ୳ࡧൣғ(ਂ͞)ΛݶఆͰ͖Δ ݱہ໘
1खޙ 2खޙ ධՁˠ 0.1 0.8 0.3 0.4
ධՁؔͷ༗ޮੑ • ύϥϝʔλͷબఆɾઃఆ͕ΩϞ • ख࡞ۀ: νΣεͰਓؒΛ͑ͨ • ػցֶश: কعͰਓؒΛ͑ͨ •
ғޟෳࡶੑͷͨΊʹ·ͱͳධՁؔΛ࡞Εͳ͔ͬ ͨ
ݪ࢝ϞϯςΧϧϩ๏ • ϥϯμϜʹऴہ·Ͱଧͭ (ϩʔϧΞτ) Λ܁Γฦͯ͠ɺউΛܭ ࢉ͢Δํ๏ ϩʔϧΞτΛ܁Γฦͯ͠ উΛܭࢉ উ 7/10
উ 3/10
ϞϯςΧϧϩ୳ࡧ (MCTS) • ݪ࢝ϞϯςΧϧϩ๏ΛධՁؔతʹ͏ • n खઌͰϩʔϧΞτ • ༿ͷউΛܭࢉ ※͞Βʹ༿ͷউʹԠͯ͡ಈతʹࢬמΓɾల։͠ɺ୳ࡧਫ਼Λ্͛Δ
ϙϦγʔؔ • f (ہ໘, ࣍ͷҰख) • ࣍ͷҰखͷࣗવ͞Λ͋ΒΘ֬͢ີؔ • عේσʔλ͔Βͷֶश͕༰қ •
ϩʔϧΞτ࣌ʹ͑Δ • ϥϯμϜʹଧͭͷͰͳ͘ɺ·ͱͳखΛଧͨͤΔ • ͨͩ͠ߴʹಈ࡞͢Δඞཁ͕͋Δ
MCTSͷڧ͞ • ϙϦγʔؔͷͳͲͰΞϚνϡΞߴஈʹඖఢ͢Δڧ͞· Ͱਐา • ϓϩʹٴͳ͍ • େہ؍ʹ༏ΕΔ • ʮڱ͘ਂ͍ಡΈʯ͕ऑ͍
• खΛ͘ಡΉͨΊ
AlphaGo͕ͬͨ͜ͱ • جຊMCTS • ༷ʑͳ • CNN(ΈࠐΈχϡʔϥϧωοτϫʔΫ) • ڧԽֶश •
ධՁؔ • ฒྻࢄΞϧΰϦζϜ • MCTS ʹͦΕΒΛΈࠐΜͩ
ୈೋ෦: AlphaGo
2ͭͷϙϦγʔؔͱ 1ͭͷධՁؔ ϩʔϧΞτϙϦγʔ ϩʔϧΞτʹ͏ ߴɾਫ਼ 4-ϙϦγʔ ୳ࡧॱংΛܾΊΔ ɾߴਫ਼ ධՁؔ ༿ͷධՁ(উ)Λܭࢉ
ϩʔϧΞτʹΑΔউͱ͠߹ΘͤΔ
ϩʔϧΞτϙϦγʔ • ϩʔϧΞτ(ϥϯμϜϓϨΠ)ʹ͏ϙϦγʔؔ • ߴੑɹʼɹਫ਼ • ਓؒͷعේ800ສہ໘͔Βֶश • ઢܗιϑτϚοΫεؔ •
2ϚΠΫϩඵ (ߴ) • عේͱͷࢦ͠खҰக: 24.2%
SLϙϦγʔ • ͷ୳ࡧॱংΛܾΊΔϙϦγʔؔ • ਫ਼ɹʼɹߴੑ • ਓؒͷعේ3000ສہ໘͔Βֶश • 13CNN(ΈࠐΈχϡʔϥϧωοτϫʔΫ) •
ը૾ೝࣝͰΑ͘ΘΕΔ • : 3ϛϦඵ • عේͱͷࢦ͠खҰக: 57%
ධՁؔ • 14CNN • SLϙϦγʔΛڧԽֶशͨ͠ͷ (RLϙϦγʔ) Λݩʹɺճؼͯ͠࡞Δ 4-ϙϦγʔ 3-ϙϦγʔ ධՁؔ
1. ڧԽֶश 2. ϥϯμϜعේੜ (3000ສہ໘) 3. ճؼ
ධՁؔͷଊ͑ํ • ϩʔϧΞτʹΑΔউܭࢉΛิ͏ͷ • ୯ମͰͦ͜·Ͱڧ͘ͳ͍ • ධՁؔͷಛ (ߟ) • ʮڱ͘ਂ͍ಡΈʯʹڧ͍
• ʮRLϙϦγʔ(ڧԽֶश݁Ռ)Λऴہ·ͰଧͨͤͨࡍͷউʯͱՁ • େہ؍͕ແ͍ • Ұຊಓ͔͠ಡ·ͳ͍ .$54ͷಛੑ େہ؍ʹ༏Εͯʮਂ͍ಡΈʯ͕ऑ͍ ͱ ͏·͘ิ͍͍͋ͬͯΔ
ڧ͞ (2015/10࣌)
ڧ͞ (2016/3 ࣌) R3500+ ͷΠɾηυϧʹউ
ࢀߟ • Mastering the game of Go with deep neural
networks and tree search (http://www.nature.com/nature/journal/v529/n7587/full/ nature16961.html) • Google AlphaGoͷΈΛཧղ͢Δ | IT Leaders (http://it.impressbm.co.jp/articles/-/13474)
ऴΘΓ