Saved in:
Bibliographic Details
Main Authors: Corrado, Nicholas E., Hanna, Josiah P.
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2311.08290
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908823607836672
author Corrado, Nicholas E.
Hanna, Josiah P.
author_facet Corrado, Nicholas E.
Hanna, Josiah P.
contents On-policy reinforcement learning (RL) algorithms are typically characterized as algorithms that perform policy updates using i.i.d. trajectories collected by the agent's current policy. However, after observing only a finite number of trajectories, such on-policy sampling may produce data that fails to match the expected on-policy data distribution. This sampling error leads to high-variance gradient estimates that yield data-inefficient on-policy learning. Recent work in the policy evaluation setting has shown that non-i.i.d., off-policy sampling can produce data with lower sampling error w.r.t. the expected on-policy distribution than on-policy sampling can produce (Zhong et. al, 2022). Motivated by this observation, we introduce an adaptive, off-policy sampling method to reduce sampling error during on-policy policy gradient RL training. Our method, Proximal Robust On-Policy Sampling (PROPS), reduces sampling error by collecting data with a behavior policy that increases the probability of sampling actions that are under-sampled w.r.t. the current policy. We empirically evaluate PROPS on continuous-action MuJoCo benchmark tasks as well as discrete-action tasks and demonstrate that (1) PROPS decreases sampling error throughout training and (2) increases the data efficiency of on-policy policy gradient algorithms.
format Preprint
id arxiv_https___arxiv_org_abs_2311_08290
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On-Policy Policy Gradient Reinforcement Learning Without On-Policy Sampling
Corrado, Nicholas E.
Hanna, Josiah P.
Machine Learning
On-policy reinforcement learning (RL) algorithms are typically characterized as algorithms that perform policy updates using i.i.d. trajectories collected by the agent's current policy. However, after observing only a finite number of trajectories, such on-policy sampling may produce data that fails to match the expected on-policy data distribution. This sampling error leads to high-variance gradient estimates that yield data-inefficient on-policy learning. Recent work in the policy evaluation setting has shown that non-i.i.d., off-policy sampling can produce data with lower sampling error w.r.t. the expected on-policy distribution than on-policy sampling can produce (Zhong et. al, 2022). Motivated by this observation, we introduce an adaptive, off-policy sampling method to reduce sampling error during on-policy policy gradient RL training. Our method, Proximal Robust On-Policy Sampling (PROPS), reduces sampling error by collecting data with a behavior policy that increases the probability of sampling actions that are under-sampled w.r.t. the current policy. We empirically evaluate PROPS on continuous-action MuJoCo benchmark tasks as well as discrete-action tasks and demonstrate that (1) PROPS decreases sampling error throughout training and (2) increases the data efficiency of on-policy policy gradient algorithms.
title On-Policy Policy Gradient Reinforcement Learning Without On-Policy Sampling
topic Machine Learning
url https://arxiv.org/abs/2311.08290