Combining AI Control Systems and Human Decision Support via Robustness and Criticality

TL;DR

This paper presents a robust AI control system integrated with human decision support, utilizing adversarial explanations and autoencoders to improve safety, robustness, and training in reinforcement learning frameworks like MuZero.

Contribution

It extends adversarial explanation methodology to advanced reinforcement learning, introduces autoencoders for saliency detection, and combines these with criticality analysis for improved human-AI collaboration.

Findings

AI control systems show increased robustness against adversarial tampering.

Adversarial explanations effectively aid human decision-making.

The integrated system enhances AI decision quality and safety in critical scenarios.

Abstract

AI-enabled capabilities are reaching the requisite level of maturity to be deployed in the real world, yet do not always make correct or safe decisions. One way of addressing these concerns is to leverage AI control systems alongside and in support of human decisions, relying on the AI control system in safe situations while calling on a human co-decider for critical situations. We extend a methodology for adversarial explanations (AE) to state-of-the-art reinforcement learning frameworks, including MuZero. Multiple improvements to the base agent architecture are proposed. We demonstrate how this technology has two applications: for intelligent decision tools and to enhance training / learning frameworks. In a decision support context, adversarial explanations help a user make the correct decision by highlighting those contextual factors that would need to change for a different AI-recommended decision. As another benefit of adversarial explanations, we show that the learned AI control system demonstrates robustness against adversarial tampering. Additionally, we supplement AE by introducing strategically similar autoencoders (SSAs) to help users identify and understand all salient factors being considered by the AI system. In a training / learning framework, this technology can improve both the AI's decisions and explanations through human interaction. Finally, to identify when AI decisions would most benefit from human oversight, we tie this combined system to our prior art on statistically verified analyses of the criticality of decisions at any point in time.