Learning Complementary Representations of the Past using Auxiliary Tasks in Partially Observable Reinforcement Learning.
Andrea BaiseroChristopher AmatoPublished in: AAMAS (2020)
Keyphrases
- reinforcement learning
- partially observable
- partially observable environments
- state space
- learning process
- markov decision processes
- partial observability
- learning algorithm
- action models
- hidden state
- dynamical systems
- partially observable domains
- learning tasks
- decision problems
- supervised learning
- function approximation
- markov decision problems
- partial observations
- infinite horizon
- reinforcement learning algorithms
- machine learning
- model free
- reward function
- belief state
- search strategies
- domain independent