pyconkr 2018 RL_Adventure : Rainbow(value based Reinforcement Learning)
?
1 like?244 views
??? ??? 2018?? ???? ??? "RL Adventure : DQN ?? Rainbow DQN??"? ?? ?????. 4?? ??? Rainbow? ?? ????? https://github.com/hongdam/pycon2018-RL_Adventure
pyconkr 2018 RL_Adventure : Rainbow(value based Reinforcement Learning)
- 2. INDEX 1. Environment 2. Before RAINBOW DDQN(Double Deep Q-Learning) Dueling DQN Multi-Step TD(Temporal Difference) PER(Prioritized Experience Replay) Noisy Network Categorical DQN(C51) 3. RAINBOW 4. RAINBOW - Code 2
- 4. 2. BEFORE RAINBOW : DOUBLE DQN 4 HTTPS://ARXIV.ORG/ABS/1509.06461
- 5. 2. BEFORE RAINBOW : DUELING DQN HTTPS://ARXIV.ORG/ABS/1511.06581 5
- 6. 2. BEFORE RAINBOW : DUELING DQN 6 HTTPS://ARXIV.ORG/ABS/1511.06581
- 7. 2. BEFORE RAINBOW : MULTI-STEP LEARNING 7
- 8. 2. BEFORE RAINBOW : PER HTTPS://ARXIV.ORG/ABS/1511.05952 8
- 9. 2. BEFORE RAINBOW : NOISY NETWORK HTTPS://ARXIV.ORG/ABS/1706.10295 9
- 10. 2. BEFORE RAINBOW : NOISY NETWORK HTTPS://ARXIV.ORG/ABS/1706.10295 10
- 11. 2. BEFORE RAINBOW : CATEGORICAL DQN(C51) HTTPS://ARXIV.ORG/PDF/1707.06887.PDF 11
- 12. 2. BEFORE RAINBOW : CATEGORICAL DQN(C51) HTTPS://ARXIV.ORG/PDF/1707.06887.PDF 12
- 14. 3. RAINBOW RAINBOW DDQN(Double Deep Q-Learning) + Dueling DQN + Multi-Step TD(Temporal Difference) + PER(Prioritized Experience Replay) + Noisy Network + Categorical DQN(C51) 14
- 15. 3. RAINBOW 15
- 17. 3. RAINBOW 17
- 18. 3. RAINBOW 18
- 19. PONG 4. RAINBOW - CODE 19
- 20. NOISY LINEAR 4. RAINBOW - CODE 20
- 21. DUELING + NOISY + C51 4. RAINBOW - CODE 21
- 22. PROJECTION STEP 4. RAINBOW - CODE 22
- 23. CROSS-ENTROPY LOSS 4. RAINBOW - CODE 23
- 24. TEST 4. RAINBOW - CODE 24