Optimising Human-AI Collaboration by Learning Convincing Explanations

TL;DR

This paper introduces Ardent, an algorithm that learns personalized explanations to improve human-AI collaboration, ensuring safety and transparency in high-stakes decision-making through adaptive, interactive learning.

Contribution

The paper presents Ardent, a novel algorithm that personalizes explanations in human-AI collaboration, enhancing decision support by learning individual preferences through interaction.

Findings

Ardent outperforms competing systems in simulations and user studies.

Personalized explanations improve decision accuracy and user trust.

Adaptive explanations lead to safer and more effective collaboration.

Abstract

Machine learning models are being increasingly deployed to take, or assist in taking, complicated and high-impact decisions, from quasi-autonomous vehicles to clinical decision support systems. This poses challenges, particularly when models have hard-to-detect failure modes and are able to take actions without oversight. In order to handle this challenge, we propose a method for a collaborative system that remains safe by having a human ultimately making decisions, while giving the model the best opportunity to convince and debate them with interpretable explanations. However, the most helpful explanation varies among individuals and may be inconsistent across stated preferences. To this end we develop an algorithm, Ardent, to efficiently learn a ranking through interaction and best assist humans complete a task. By utilising a collaborative approach, we can ensure safety and improve performance while addressing transparency and accountability concerns. Ardent enables efficient and effective decision-making by adapting to individual preferences for explanations, which we validate through extensive simulations alongside a user study involving a challenging image classification task, demonstrating consistent improvement over competing systems.