2026-09-19T17:52:15Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1518272023-08-24T03:49:19Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Sontag, David
Mozannar, Hussein
Massachusetts Institute of Technology. Institute for Data, Systems, and Society
2023-08-23T16:11:39Z
2023-08-23T16:11:39Z
2023-06
2023-07-24T18:01:48.706Z
https://hdl.handle.net/1721.1/151827
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 thesis, 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.
S.M.
Massachusetts Institute of Technology
Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)
Copyright retained by author(s)
https://creativecommons.org/licenses/by-sa/4.0/
Consistent Estimators for Learning to Defer to an Expert
Thesis
application/pdf
Master
Master of Science in Social and Engineering Systems
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="26e900d7-33d9-43e7-9de0-5879ec4e94ff">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Consistent Estimators for Learning to Defer to an Expert</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2023-06</PublicationDate>
<Authors>
<Author>
<DisplayName>Mozannar, Hussein</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://creativecommons.org/licenses/by-sa/4.0/</License>
<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 thesis, 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.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>