DeepRacer for learning RL

2019.4.6 Developers.IO at OKAYAMA.

Transcript

1. 4 D 29 26 1 . 0 I 1
2. & .-*  (2  ,0'/4"# 51 
   83;7 +) !&%$9( 6:  Attention
3. 3 
    #cmdevio2019
4. 4 os t m ( L @S g E b i _d L rMI D @ E ( ( ( ) ( e a n k AWS E
5. 5  
    DeepRacer 
6. 6  
    D
7. 7    
    ) (
8. 8 …
9. 9 DeepRacer
   
   
   
   4 D 26 9 01 .
10. 10 DeepRacer
    
    
    
    A A A
11. 11
    
       
        1 2 3
12. 12 DeepRacer
    
13. 13  
     
14. 14 DeepRacer 1/18
    
    
    
          3D
     
    
    AWS DeepRacer League
15. 15 DeepRacer
     https://aws.amazon.com/jp/deepracer/
16. 16 DeepRacer
     ! &%$ +)*2 1 '/
    *2     (#-, 0.  "
17. 17 3D
    
    
    AWS RoboMaker Robot Operating System (ROS) Gazebo rqt
18. 18 AWS DeepRacer League ⁻ 0 1 : 9 A ⁻ 9 2 R ⁻  
     ⁻ D I ⁻ 1 2 ⁻   https://aws.amazon.com/jp/deepracer/league/
19. 19 
    
20. 20   (Artificial Intelligence, AI)  (Machine Learning, ML) 
       NeuralNetwork DeepLearning   
21. 21 
    
22. 22 = 1    (      ) ( (
23. 23 L N - ) ( - D Q
24. 24 DeepRacer     Cliped PPO PPO (Proximal Policy Optimization) OpenAI2017 
25. 25 
         ( ( ) )
26. 26 
     1
27. 27 
     ) () (
28. 28 DeepRacer
      
29. 29 DeepRacer
     + + +
30. 30 DeepRacer
    
31. 31 DeepRacer 
    …
32. 32 DeepRacer 
     
    
33. 33 orz
34. 34 DeepRacer 
    + + +
35. 35 DeepRacer 
    D A D
36. 36 $
    
      ' + (# 
      &! %"
37. 37  
     ( ) ) https://docs.aws.amazon.com/ja_jp /deepracer/latest/developerguide/ deepracer-train-models-define- reward-function.html
38. 38 
      
     
39. 39   
        ⁻ 10 ⁻ :
40. 40 SageMaker RL + RoboMaker
41. 41 SageMeker RLRoboMakerGA
42. 42 SageMaker “RL”  ⁻ ⁻ M ⁻ M M ⁻ M  
    ⁻ ⁻ ⁻ J M S
43. 43 DeepRacer ) D ( ) ) ( )
44. 44 SageMaker 
          
       https://dev.classmethod.jp/machine -learning/sagemaker-robomaker- deepracer-sample/
45. 45
         $#   "
        ! https://github.com/awslabs/amazon-sagemaker-examples
46. 46 Jupyter
       !
      
47. 47 
     ( ( ) 
48. 48   
49. 49  2   1  
    
50. 50  2   2  
         
    
51. 51   (. ( )( 
      
52. 52 $2# " /1(+ $2#,- !  https://docs.aws.amazon.com/ja_jp/deepracer/latest/developerguide/deepracer -iteratively-enhance-reward-functions.html 
    *  $2%) '.0& )
53. 53 
     
     Best Practices when training with PPO (Unity Technologies) https://github.com/Unity-Technologies/ml-agents/blob/master/docs/Training-PPO.md   
    
        
54. 54 DeepRacer "% 
     
      !  
    "%  
     
     
     $#  
    
     
     !  
    
    !
55. 55 DeepRacer
    
56. 56 DeepRacer
     
       
57. 57 DeepRacer
     
    
58. 58 DeepRacer
        
    
59. 59 DeepRacer
       
      
60. 60 
    
61. 61 • g • + + • M D c • R S D a k • b • LL e
62. 62 DeepRacer