Revisiting the Softmax Bellman Operator: New Benefits and New Perspective
This addresses the challenge of improving policy performance in reinforcement learning for practitioners, though it is incremental as it revisits and refines an existing operator.
The paper tackles the problem of the softmax Bellman operator in reinforcement learning, which is often seen as sub-optimal, and finds that it leads to superior policies in practice, even outperforming double Q-learning in Deep Q-learning setups.
The impact of softmax on the value function itself in reinforcement learning (RL) is often viewed as problematic because it leads to sub-optimal value (or Q) functions and interferes with the contraction properties of the Bellman operator. Surprisingly, despite these concerns, and independent of its effect on exploration, the softmax Bellman operator when combined with Deep Q-learning, leads to Q-functions with superior policies in practice, even outperforming its double Q-learning counterpart. To better understand how and why this occurs, we revisit theoretical properties of the softmax Bellman operator, and prove that $(i)$ it converges to the standard Bellman operator exponentially fast in the inverse temperature parameter, and $(ii)$ the distance of its Q function from the optimal one can be bounded. These alone do not explain its superior performance, so we also show that the softmax operator can reduce the overestimation error, which may give some insight into why a sub-optimal operator leads to better performance in the presence of value function approximation. A comparison among different Bellman operators is then presented, showing the trade-offs when selecting them.