Expert-Agnostic Learning to Defer

2025-02-14

AI Summary

This paper addresses the challenge of poor generalization in Learning to Defer (L2D) to unseen human experts. We propose the first expert-agnostic Bayesian modeling framework for L2D, which explicitly models predictive uncertainty, abstracts expert behavior via latent representations, and decouples decision logic from expert identity—enabling principled incorporation of prior knowledge and substantially reducing reliance on expert-labeled data. Evaluated on six benchmark datasets, our method achieves average improvements of 1–16% on seen experts and 4–28% on unseen experts over state-of-the-art approaches. Our core contribution is the first L2D framework achieving robust generalization to previously unobserved experts, simultaneously ensuring low annotation cost and high decision reliability—establishing a new paradigm for trustworthy human-AI collaborative autonomous systems.

Technology Category

• Reasoning under Uncertainty: Sequential Decision Making
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Learning to Defer (L2D) learns autonomous systems to independently manage straightforward cases, while deferring uncertain cases to human experts. Recent advancements in this field have introduced features enabling flexibility to unseen experts at test-time, but we find these approaches have significant limitations. To address these, we introduce EA-L2D: Expert-Agnostic Learning to Defer, a novel L2D framework that leverages a Bayesian approach to model expert behaviour in an expert-agnostic manner, facilitating optimal deferral decisions. EA-L2D offers several critical improvements over prior methods, including the ability to incorporate prior knowledge about experts, a reduced reliance on expert-annotated data, and robust performance when deferring to experts with expertise not seen during training. Evaluating on CIFAR-10, HAM10000, German Traffic Lights, Breast Ultrasound, Axial Organ Slices, and Blood Cell MNIST, we observe performance gains over the next state-of-the-art of 1-16% for seen experts and 4-28% for unseen experts in settings with high expert diversity.

Problem

• Improves expert-agnostic decision deferral
• Reduces reliance on expert-annotated data
• Enhances performance with diverse expert expertise

Innovation

• Bayesian expert behavior modeling
• Reduced expert-annotated data dependency
• Robust deferral to unseen experts