Sample-Efficient Reinforcement Learning with Maximum Entropy Mellowmax Episodic Control
arXiv:1911.09615
Abstract
Deep networks have enabled reinforcement learning to scale to more complex and challenging domains, but these methods typically require large quantities of training data. An alternative is to use sample-efficient episodic control methods: neuro-inspired algorithms which use non-/semi-parametric models that predict values based on storing and retrieving previously experienced transitions. One way to further improve the sample efficiency of these approaches is to use more principled exploration strategies. In this work, we therefore propose maximum entropy mellowmax episodic control (MEMEC), which samples actions according to a Boltzmann policy with a state-dependent temperature. We demonstrate that MEMEC outperforms other uncertainty- and softmax-based exploration methods on classic reinforcement learning environments and Atari games, achieving both more rapid learning and higher final rewards.
Workshop on Biological and Artificial Reinforcement Learning, NeurIPS 2019
References in corpus (10)
- A Brief Survey of Deep Reinforcement Learning
- Dueling Network Architectures for Deep Reinforcement Learning
- (More) Efficient Reinforcement Learning via Posterior Sampling
- Bridging the Gap Between Value and Policy Based Reinforcement Learning
- Model-Free Episodic Control
- A unified view of entropy-regularized Markov decision processes
- Neural Episodic Control
- UCB Exploration via Q-Ensembles
- A Laplacian Framework for Option Discovery in Reinforcement Learning
- A short variational proof of equivalence between policy gradients and soft Q learning