Smoothed Dual Embedding Control.

We revisit the Bellman optimality equation with Nesterov's smoothing technique and provide a unique saddle-point optimization perspective of the policy optimization problem in reinforcement learning based on Fenchel duality. A new reinforcement learning algorithm, called Smoothed Dual Embedding Control or SDEC, is derived to solve the saddle-point reformulation with arbitrary learnable function approximator. The algorithm bypasses the policy evaluation step in the policy optimization from a principled scheme and is extensible to integrate with multi-step bootstrapping and eligibility traces. We provide a PAC-learning bound on the number of samples needed from one single off-policy sample path, and also characterize the convergence of the algorithm. Finally, we show the algorithm compares favorably to the state-of-the-art baselines on several benchmark control problems.