Upgrade to Pro — share decks privately, control downloads, hide ads and more …

強化学習で何ができるようになるのか / What can you do with reinforcement learning

kodaisu
October 02, 2018

強化学習で何ができるようになるのか / What can you do with reinforcement learning

Platinam Data Meetup Vol.2(https://brainpad-meetup.connpass.com/event/100568/)
の5分LT枠で発表した際の資料です。

kodaisu

October 02, 2018
Tweet

Other Decks in Technology

Transcript

  1. .
    5

    View Slide





  2. View Slide


  3. 5 5 5

    View Slide



  4. .5 3 2
    4
    12

    View Slide



  5. View Slide

  6. 5
    S (
    ) (
    ) ) ( D R

    View Slide


  7. a
    d D
    S e

    View Slide


  8. )
    (

    (
    (

    View Slide


  9. )

    )
    ( )
    (

    View Slide



  10. View Slide

  11. L MSN
    e ) - ) ( ) hi
    1 0 1 0
    kt N gl L MSN
    G a
    G Aru o
    L MSN
    c
    w

    View Slide



  12. 1
    4
    -
    4

    View Slide

  13. !

    -
    A


    View Slide




  14. A
    d % A
    e
    N a - A
    r A
    S w A
    N

    View Slide

  15. e

    #" !
    W T- W
    - ) LC b ) -
    ( W W
    V R W ) W

    View Slide





  16. View Slide



  17. A 7HDSPLPJ SDPTIHSDENH /S KLUH U SHT IQS DNDENH 4ODJH H QJPLULQP
    KUURT- DS L QSJ R I R I
    A HLPIQS HOHPU 7HDSPLPJ LP 94
    KUURT- OH L O QO . LNL SHLPIQS HOHPU NHDSPLPJ LP PLRT , DE DE
    A @TLPJ 1HHR 7HDSPLPJ UQ 0QPUSQN :RULOL[DULQP 3 RHSRDSDOHUHST
    KUURT- DS L QSJ R I R I
    A 2IIL LHPU 9H SDN /S KLUH U SH HDS K LD DSDOHUHS KDSLPJ
    KUURT- DS L QSJ R I R I
    A HLPIQS HOHPU 7HDSPLPJ LP QEQUL T- / S H
    KUURT- LDT LPIQSODULM U DSOTUD U H RNQD T ENL DULQPT 6QEHSC45 C R I
    A 1HHR HLPIQS HOHPU 7HDSPLPJ IQS QEQUL 8DPLR NDULQP LUK /T P KSQPQ T :II QNL
    @R DUHT
    KUURT- DS L QSJ R I R I

    View Slide



  18. 3 AM=M?D ?=H 7 A?O 2AOA?O OD 2AAL 8A M?AIA O 5A=M C
    DOOL - I=OCA L? C OD AOA?O L
    8A M?A 6= =CAIA O OD 2AAL 8A M?AIA O 5A=M C
    DOOL - H =?I MC ? O=O ? I/ .
    0 8A M?AIA O 5A=M C 0LLM =?D O 7 H A :A 9S OAI 0 O ? C M=O
    DOOL- M= CAM O= A M= L=LAM 41219 , L
    , 8A M?AIA O HA=M C = A I HO =CA O S OAI M AO M OM= ? C =H ? OM H
    DOOL - AAARLH MA AAA MC ? IA O

    View Slide