Neural Episodic Control
Alexander Pritzel, Benigno Uría, Srinivasan, Sriram et al. · arXiv (Cornell University) · 2017
Deep reinforcement learning methods attain super-human performance in a wide range of environments. Such methods are grossly inefficient, often taking orders of magnitudes more data than humans to achieve reasonable performance. We propose Neural Episodic Control: a deep reinforcement learning agent that is able to rapidly assimilate new experiences and act upon them. Our agent uses a semi-tabular representation of the value function: a buffer of past experience containing slowly changing state representations and rapidly updated estimates of the value function. We show across a wide range of environments that our agent learns significantly faster than other state-of-the-art, general purpose deep reinforcement learning agents.
Cited by 78 works. doi:10.48550/arxiv.1703.01988 Open-access PDF
Authors
References
- Long Short-Term Memory
- Reinforcement Learning: An Introduction
- Human-level control through deep reinforcement learning
- Multidimensional binary search trees used for associative searching
- Mastering the game of Go with deep neural networks and tree search
- Learning to Predict by the Methods of Temporal Differences
- Learning from Delayed Rewards
- Introduction to Reinforcement Learning
- Stochastic Backpropagation and Approximate Inference in Deep Generative Models
- Deep Reinforcement Learning with Double Q-Learning