Query The Agent: Improving sample efficiency through epistemic uncertainty estimation

TL;DR

This paper introduces Query The Agent (QTA), an algorithm that enhances sample efficiency in goal-conditioned reinforcement learning by estimating epistemic uncertainty across the state space and focusing on uncertain areas.

Contribution

The paper presents a novel algorithm, QTA, utilizing Predictive Uncertainty Networks (PUN) to estimate epistemic uncertainty and improve sample efficiency in reinforcement learning.

Findings

QTA significantly outperforms existing methods in sample efficiency.

PUN effectively estimates epistemic uncertainty across states.

QTA accelerates value function learning through targeted exploration.

Abstract

Curricula for goal-conditioned reinforcement learning agents typically rely on poor estimates of the agent's epistemic uncertainty or fail to consider the agents' epistemic uncertainty altogether, resulting in poor sample efficiency. We propose a novel algorithm, Query The Agent (QTA), which significantly improves sample efficiency by estimating the agent's epistemic uncertainty throughout the state space and setting goals in highly uncertain areas. Encouraging the agent to collect data in highly uncertain states allows the agent to improve its estimation of the value function rapidly. QTA utilizes a novel technique for estimating epistemic uncertainty, Predictive Uncertainty Networks (PUN), to allow QTA to assess the agent's uncertainty in all previously observed states. We demonstrate that QTA offers decisive sample efficiency improvements over preexisting methods.