Regression with Multi-Expert Deferral

TL;DR

This paper introduces a novel framework for regression with multiple experts, providing new surrogate loss functions, theoretical guarantees, and algorithms that improve upon existing methods for deferral in regression tasks.

Contribution

It develops a comprehensive regression with deferral framework, including new surrogate losses, non-asymptotic consistency bounds, and algorithms applicable to multiple experts and various cost structures.

Findings

Proposed surrogate losses supported by $H$-consistency bounds.

Framework applies to multiple experts and various loss functions.

Experimental results demonstrate effectiveness of new algorithms.

Abstract

Learning to defer with multiple experts is a framework where the learner can choose to defer the prediction to several experts. While this problem has received significant attention in classification contexts, it presents unique challenges in regression due to the infinite and continuous nature of the label space. In this work, we introduce a novel framework of regression with deferral, which involves deferring the prediction to multiple experts. We present a comprehensive analysis for both the single-stage scenario, where there is simultaneous learning of predictor and deferral functions, and the two-stage scenario, which involves a pre-trained predictor with a learned deferral function. We introduce new surrogate loss functions for both scenarios and prove that they are supported by $H$-consistency bounds. These bounds provide consistency guarantees that are stronger than Bayes consistency, as they are non-asymptotic and hypothesis set-specific. Our framework is versatile, applying to multiple experts, accommodating any bounded regression losses, addressing both instance-dependent and label-dependent costs, and supporting both single-stage and two-stage methods. A by-product is that our single-stage formulation includes the recent regression with abstention framework (Cheng et al., 2023) as a special case, where only a single expert, the squared loss and a label-independent cost are considered. Minimizing our proposed loss functions directly leads to novel algorithms for regression with deferral. We report the results of extensive experiments showing the effectiveness of our proposed algorithms.