Optimal oracle inequalities for solving projected fixed-point equations
Mou, Wenlong, Pananjady, Ashwin, Wainwright, Martin J.
Linear fixed point equations in Hilbert spaces arise in a variety of settings, including reinforcement learning, and computational methods for solving differential and integral equations. We study methods that use a collection of random observations to compute approximate solutions by searching over a known low-dimensional subspace of the Hilbert space. First, we prove an instance-dependent upper bound on the mean-squared error for a linear stochastic approximation scheme that exploits Polyak--Ruppert averaging. This bound consists of two terms: an approximation error term with an instance-dependent approximation factor, and a statistical error term that captures the instance-specific complexity of the noise when projected onto the low-dimensional subspace. Using information theoretic methods, we also establish lower bounds showing that both of these terms cannot be improved, again in an instance-dependent sense. A concrete consequence of our characterization is that the optimal approximation factor in this problem can be much larger than a universal constant. We show how our results precisely characterize the error of a class of temporal difference learning methods for the policy evaluation problem with linear function approximation, establishing their optimality.
Dec-9-2020
- Country:
- North America > United States
- Massachusetts > Middlesex County
- Belmont (0.04)
- California > Alameda County
- Berkeley (0.04)
- Massachusetts > Middlesex County
- Europe > United Kingdom
- England
- Cambridgeshire > Cambridge (0.04)
- Oxfordshire > Oxford (0.04)
- England
- Asia > Middle East
- Jordan (0.04)
- North America > United States
- Genre:
- Research Report > New Finding (0.66)
- Technology: