Learning to Complement and to Defer to Multiple Users

TL;DR

This paper introduces LECODU, a unified approach that combines learning to complement and defer to multiple users in human-AI classification, optimizing accuracy and user collaboration costs.

Contribution

LECODU is the first method to unify complementing and deferring strategies with user number estimation in HAI-CC, improving performance over existing methods.

Findings

LECODU outperforms state-of-the-art HAI-CC methods.

LECODU maintains robustness with noisy, unreliable users.

LECODU enhances decision accuracy and reduces collaboration costs.

Abstract

With the development of Human-AI Collaboration in Classification (HAI-CC), integrating users and AI predictions becomes challenging due to the complex decision-making process. This process has three options: 1) AI autonomously classifies, 2) learning to complement, where AI collaborates with users, and 3) learning to defer, where AI defers to users. Despite their interconnected nature, these options have been studied in isolation rather than as components of a unified system. In this paper, we address this weakness with the novel HAI-CC methodology, called Learning to Complement and to Defer to Multiple Users (LECODU). LECODU not only combines learning to complement and learning to defer strategies, but it also incorporates an estimation of the optimal number of users to engage in the decision process. The training of LECODU maximises classification accuracy and minimises collaboration costs associated with user involvement. Comprehensive evaluations across real-world and synthesized datasets demonstrate LECODU's superior performance compared to state-of-the-art HAI-CC methods. Remarkably, even when relying on unreliable users with high rates of label noise, LECODU exhibits significant improvement over both human decision-makers alone and AI alone.