Timezone: »

Bayesian Q-learning With Imperfect Expert Demonstrations
Fengdi Che · Xiru Zhu · Doina Precup · David Meger · Gregory Dudek
Event URL: https://openreview.net/forum?id=rIq73puh9- »

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. We can achieve better results with the proposed methods than Deep Q-learning from Demonstrations (Hester et al., 2017) in most environments.

Author Information

Fengdi Che (University of Alberta)
Xiru Zhu
Doina Precup (McGill University / Mila / DeepMind Montreal)
David Meger (McGill University)
Gregory Dudek (McGill University & Samsung Research)

More from the Same Authors