Uniform-PAC Bounds for Reinforcement Learning with Linear Function Approximation
This work addresses a fundamental limitation in RL theory for researchers and practitioners, providing a more robust convergence guarantee, though it is incremental in advancing theoretical bounds rather than introducing a new paradigm.
The paper tackles the problem of reinforcement learning with linear function approximation by proposing a new algorithm called FLUTE, which achieves uniform-PAC convergence to the optimal policy with high probability, a stronger guarantee than existing methods that only offer high-probability regret or PAC bounds.
We study reinforcement learning (RL) with linear function approximation. Existing algorithms for this problem only have high-probability regret and/or Probably Approximately Correct (PAC) sample complexity guarantees, which cannot guarantee the convergence to the optimal policy. In this paper, in order to overcome the limitation of existing algorithms, we propose a new algorithm called FLUTE, which enjoys uniform-PAC convergence to the optimal policy with high probability. The uniform-PAC guarantee is the strongest possible guarantee for reinforcement learning in the literature, which can directly imply both PAC and high probability regret bounds, making our algorithm superior to all existing algorithms with linear function approximation. At the core of our algorithm is a novel minimax value function estimator and a multi-level partition scheme to select the training samples from historical observations. Both of these techniques are new and of independent interest.