#Exploration: A Study of Count-Based Exploration for Deep Reinforcement Learning

Count-based exploration algorithms are known to perform near-optimally when used in conjunction with tabular reinforcement learning (RL) methods for solving small discrete Markov decision processes (MDPs). It is generally thought that count-based methods cannot be applied in high-dimensional state spaces, since most states will only occur once. Recent deep RL exploration strategies are able to deal with high-dimensional continuous state spaces through complex heuristics, often relying on optimism in the face of uncertainty or intrinsic motivation. In this work, we describe a surprising finding: a simple generalization of the classic count-based approach can reach near state-of-the-art performance on various high-dimensional and/or continuous deep RL benchmarks. States are mapped to hash codes, which allows to count their occurrences with a hash table. These counts are then used to compute a reward bonus according to the classic count-based exploration theory. We find that simple hash functions can achieve surprisingly good results on many challenging tasks. Furthermore, we show that a domain-dependent learned hash code may further improve these results. Detailed analysis reveals important aspects of a good hash function: 1) having appropriate granularity and 2) encoding information relevant to solving the MDP. This exploration strategy achieves near state-of-the-art performance on both continuous control tasks and Atari 2600 games, hence providing a simple yet powerful baseline for solving MDPs that require considerable exploration.

PDF Abstract NeurIPS 2017 PDF NeurIPS 2017 Abstract

Results from the Paper

Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Atari Games Atari 2600 Freeway TRPO-hash Score 34.0 # 1
Atari Games Atari 2600 Frostbite TRPO-hash Score 5214.0 # 14
Atari Games Atari 2600 Montezuma's Revenge TRPO-hash Score 75 # 26
Atari Games Atari 2600 Venture TRPO-hash Score 445.0 # 20


No methods listed for this paper. Add relevant methods here