Safety Margins for Reinforcement Learning

TL;DR

This paper introduces a method to quantify and monitor the safety margins of reinforcement learning agents by comparing proxy criticality metrics to true criticality, enabling real-time detection of unsafe situations.

Contribution

It proposes a robust way to define and compute safety margins for RL agents using proxy metrics, facilitating real-time safety monitoring.

Findings

Safety margins decrease near failure states

Proxy metrics effectively approximate true criticality

Method applied to Atari RL policies

Abstract

Any autonomous controller will be unsafe in some situations. The ability to quantitatively identify when these unsafe situations are about to occur is crucial for drawing timely human oversight in, e.g., freight transportation applications. In this work, we demonstrate that the true criticality of an agent's situation can be robustly defined as the mean reduction in reward given some number of random actions. Proxy criticality metrics that are computable in real-time (i.e., without actually simulating the effects of random actions) can be compared to the true criticality, and we show how to leverage these proxy metrics to generate safety margins, which directly tie the consequences of potentially incorrect actions to an anticipated loss in overall performance. We evaluate our approach on learned policies from APE-X and A3C within an Atari environment, and demonstrate how safety margins decrease as agents approach failure states. The integration of safety margins into programs for monitoring deployed agents allows for the real-time identification of potentially catastrophic situations.