ConQUR: Mitigating Delusional Bias in Deep Q-learning
Su, Andy, Ooi, Jayden, Lu, Tyler, Schuurmans, Dale, Boutilier, Craig
–arXiv.org Artificial Intelligence
Delusional bias is a fundamental source of error in approximate Q-learning. To date, the only techniques that explicitly address delusion require comprehensive search using tabular value estimates. In this paper, we develop efficient methods to mitigate delusional bias by training Q-approximators with labels that are "consistent" with the underlying greedy policy class. We introduce a simple penalization scheme that encourages Q-labels used across training batches to remain (jointly) consistent with the expressible policy class. We also propose a search framework that allows multiple Q-approximators to be generated and tracked, thus mitigating the effect of premature (implicit) policy commitments. Experimental results demonstrate that these methods can improve the performance of Q-learning in a variety of Atari games, sometimes dramatically.
arXiv.org Artificial Intelligence
Feb-27-2020
- Country:
- North America
- United States
- Massachusetts > Middlesex County
- Cambridge (0.14)
- California > Santa Clara County
- Mountain View (0.04)
- Massachusetts > Middlesex County
- Canada
- Quebec > Montreal (0.04)
- Ontario > Toronto (0.04)
- British Columbia > Metro Vancouver Regional District
- Vancouver (0.04)
- Alberta > Census Division No. 11
- Edmonton Metropolitan Region > Edmonton (0.04)
- United States
- Europe
- United Kingdom > England
- Cambridgeshire > Cambridge (0.04)
- Portugal > Porto
- Porto (0.04)
- United Kingdom > England
- Asia > Middle East
- Israel > Haifa District > Haifa (0.04)
- Africa > Ethiopia
- Addis Ababa > Addis Ababa (0.04)
- North America
- Genre:
- Research Report > New Finding (0.66)
- Industry:
- Leisure & Entertainment > Games > Computer Games (0.54)
- Technology: