Model-free Low-Rank Reinforcement Learning via Leveraged Entry-wise Matrix Estimation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Stojanovic, Stefan, Jedra, Yassir, Proutiere, Alexandre
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929585460871168
author Stojanovic, Stefan
Jedra, Yassir
Proutiere, Alexandre
author_facet Stojanovic, Stefan
Jedra, Yassir
Proutiere, Alexandre
contents We consider the problem of learning an $\varepsilon$-optimal policy in controlled dynamical systems with low-rank latent structure. For this problem, we present LoRa-PI (Low-Rank Policy Iteration), a model-free learning algorithm alternating between policy improvement and policy evaluation steps. In the latter, the algorithm estimates the low-rank matrix corresponding to the (state, action) value function of the current policy using the following two-phase procedure. The entries of the matrix are first sampled uniformly at random to estimate, via a spectral method, the leverage scores of its rows and columns. These scores are then used to extract a few important rows and columns whose entries are further sampled. The algorithm exploits these new samples to complete the matrix estimation using a CUR-like method. For this leveraged matrix estimation procedure, we establish entry-wise guarantees that remarkably, do not depend on the coherence of the matrix but only on its spikiness. These guarantees imply that LoRa-PI learns an $\varepsilon$-optimal policy using $\widetilde{O}({S+A\over \mathrm{poly}(1-γ)\varepsilon^2})$ samples where $S$ (resp. $A$) denotes the number of states (resp. actions) and $γ$ the discount factor. Our algorithm achieves this order-optimal (in $S$, $A$ and $\varepsilon$) sample complexity under milder conditions than those assumed in previously proposed approaches.
format Preprint
id arxiv_https___arxiv_org_abs_2410_23434
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Model-free Low-Rank Reinforcement Learning via Leveraged Entry-wise Matrix Estimation
Stojanovic, Stefan
Jedra, Yassir
Proutiere, Alexandre
Machine Learning
We consider the problem of learning an $\varepsilon$-optimal policy in controlled dynamical systems with low-rank latent structure. For this problem, we present LoRa-PI (Low-Rank Policy Iteration), a model-free learning algorithm alternating between policy improvement and policy evaluation steps. In the latter, the algorithm estimates the low-rank matrix corresponding to the (state, action) value function of the current policy using the following two-phase procedure. The entries of the matrix are first sampled uniformly at random to estimate, via a spectral method, the leverage scores of its rows and columns. These scores are then used to extract a few important rows and columns whose entries are further sampled. The algorithm exploits these new samples to complete the matrix estimation using a CUR-like method. For this leveraged matrix estimation procedure, we establish entry-wise guarantees that remarkably, do not depend on the coherence of the matrix but only on its spikiness. These guarantees imply that LoRa-PI learns an $\varepsilon$-optimal policy using $\widetilde{O}({S+A\over \mathrm{poly}(1-γ)\varepsilon^2})$ samples where $S$ (resp. $A$) denotes the number of states (resp. actions) and $γ$ the discount factor. Our algorithm achieves this order-optimal (in $S$, $A$ and $\varepsilon$) sample complexity under milder conditions than those assumed in previously proposed approaches.
title Model-free Low-Rank Reinforcement Learning via Leveraged Entry-wise Matrix Estimation
topic Machine Learning
url https://arxiv.org/abs/2410.23434