Correcting Momentum in Temporal Difference Learning
This addresses a specific optimization issue in deep reinforcement learning for researchers and practitioners, though it is incremental as it modifies an existing technique.
The paper tackled the problem of momentum in Temporal Difference (TD) learning causing stale gradients due to bootstrapping, and proposed a correction term that improved sample efficiency in policy evaluation, with concrete gains shown in experiments.
A common optimization tool used in deep reinforcement learning is momentum, which consists in accumulating and discounting past gradients, reapplying them at each iteration. We argue that, unlike in supervised learning, momentum in Temporal Difference (TD) learning accumulates gradients that become doubly stale: not only does the gradient of the loss change due to parameter updates, the loss itself changes due to bootstrapping. We first show that this phenomenon exists, and then propose a first-order correction term to momentum. We show that this correction term improves sample efficiency in policy evaluation by correcting target value drift. An important insight of this work is that deep RL methods are not always best served by directly importing techniques from the supervised setting.