Bayesian Q-learning With Imperfect Expert Demonstrations
This work addresses data efficiency issues in reinforcement learning for practitioners, but it is incremental as it builds on existing demonstration-guided methods.
The paper tackles the problem of overusing expert demonstrations in reinforcement learning by proposing a novel algorithm that uses limited, imperfect expert data to speed up Q-learning, achieving better results than Deep Q-learning from Demonstrations in most environments tested.
Guided exploration with expert demonstrations improves data efficiency for reinforcement learning, but current algorithms often overuse expert information. We propose a novel algorithm to speed up Q-learning with the help of a limited amount of imperfect expert demonstrations. The algorithm avoids excessive reliance on expert data by relaxing the optimal expert assumption and gradually reducing the usage of uninformative expert data. Experimentally, we evaluate our approach on a sparse-reward chain environment and six more complicated Atari games with delayed rewards. With the proposed methods, we can achieve better results than Deep Q-learning from Demonstrations (Hester et al., 2017) in most environments.