Prioritized Experience Replay

arxiv:1511.05952

Original source · Paper JSON

Reinforcement learning · 2015-11-18

Cites (2) · Cited by (15)

Complete cleaned Markdown · Source Markdown including bibliography

Read in parts

1. Part 1
2. Part 2
3. Part 3
4. Part 4

Contents

• Prioritized Experience Replay
• Abstract
• 1 Introduction
• 2 Background
• 3 Prioritized Replay
• 3.1 A Motivating Example
• 3.2 Prioritizing with TD-error
• 3.3 Stochastic Prioritization
• 3.4 Annealing the Bias
• 4 Atari Experiments
• 5 Discussion
• 6 Extensions
• 7 Conclusion
• Acknowledgments
• Appendix A Prioritization Variants
• Appendix B Experimental Details
• B.1 Blind Cliffwalk
• B.2 Atari Experiments
• B.2.1 Implementation Details
• Rank-based prioritization
• Proportional prioritization
• B.2.2 Hyperparameters
• B.2.3 Evaluation
• B.3 Class-imbalanced MNIST
• B.3.1 Dataset Setup
• B.3.2 Training Setup
• Instructions for reporting errors