Distributional Policy Optimization: An Alternative Approach for Continuous Control
This addresses a fundamental limitation in reinforcement learning for continuous control, offering a novel alternative to policy gradient methods.
The paper tackles the problem of policy gradient methods in continuous control being limited to parametric distributions, which leads to sub-optimal convergence, by proposing a distributional framework that represents arbitrary distributions without requiring knowledge of the underlying probability distribution, resulting in performance that is comparable to or surpasses state-of-the-art baselines.
We identify a fundamental problem in policy gradient-based methods in continuous control. As policy gradient methods require the agent's underlying probability distribution, they limit policy representation to parametric distribution classes. We show that optimizing over such sets results in local movement in the action space and thus convergence to sub-optimal solutions. We suggest a novel distributional framework, able to represent arbitrary distribution functions over the continuous action space. Using this framework, we construct a generative scheme, trained using an off-policy actor-critic paradigm, which we call the Generative Actor Critic (GAC). Compared to policy gradient methods, GAC does not require knowledge of the underlying probability distribution, thereby overcoming these limitations. Empirical evaluation shows that our approach is comparable and often surpasses current state-of-the-art baselines in continuous domains.