Current browse context:
cs.LG
Change to browse by:
References & Citations
Computer Science > Machine Learning
Title: Consistent Estimators for Learning to Defer to an Expert
(Submitted on 2 Jun 2020 (v1), revised 4 Oct 2020 (this version, v2), latest version 25 Jan 2021 (v3))
Abstract: Learning algorithms are often used in conjunction with expert decision makers in practical scenarios, however this fact is largely ignored when designing these algorithms. In this paper we explore how to learn predictors that can either predict or choose to defer the decision to a downstream expert. Given only samples of the expert's decisions, we give a procedure based on learning a classifier and a rejector and analyze it theoretically. Our approach is based on a novel reduction to cost sensitive learning where we give a consistent surrogate loss for cost sensitive learning that generalizes the cross entropy loss. We show the effectiveness of our approach on a variety of experimental tasks.
Submission history
From: Hussein Mozannar [view email][v1] Tue, 2 Jun 2020 18:21:38 GMT (131kb,D)
[v2] Sun, 4 Oct 2020 01:43:06 GMT (669kb,D)
[v3] Mon, 25 Jan 2021 01:43:28 GMT (671kb,D)
Link back to: arXiv, form interface, contact.