Maximum Entropy Gain Exploration for Long Horizon Multi-goal Reinforcement Learning.
Silviu PitisHarris ChanStephen ZhaoBradly C. StadieJimmy BaPublished in: ICML (2020)
Keyphrases
- maximum entropy
- reinforcement learning
- maximum entropy principle
- markov models
- random fields
- active exploration
- iterative scaling
- conditional random fields
- probabilistic logic
- minimum cross entropy
- principle of maximum entropy
- transformation based learning
- maximum entropy model
- bregman divergences
- segmentation method
- optimal policy
- ground truth
- state space
- learning algorithm
- machine learning