Generalized Strategic Classification and the Case of Aligned Incentives

Sagi Levanon, Nir Rosenfeld
Proceedings of the 39th International Conference on Machine Learning, PMLR 162:12593-12618, 2022.

Abstract

Strategic classification studies learning in settings where self-interested users can strategically modify their features to obtain favorable predictive outcomes. A key working assumption, however, is that “favorable” always means “positive”; this may be appropriate in some applications (e.g., loan approval), but reduces to a fairly narrow view of what user interests can be. In this work we argue for a broader perspective on what accounts for strategic user behavior, and propose and study a flexible model of generalized strategic classification. Our generalized model subsumes most current models but includes other novel settings; among these, we identify and target one intriguing sub-class of problems in which the interests of users and the system are aligned. This setting reveals a surprising fact: that standard max-margin losses are ill-suited for strategic inputs. Returning to our fully generalized model, we propose a novel max-margin framework for strategic learning that is practical and effective, and which we analyze theoretically. We conclude with a set of experiments that empirically demonstrate the utility of our approach.

Cite this Paper


BibTeX
@InProceedings{pmlr-v162-levanon22a, title = {Generalized Strategic Classification and the Case of Aligned Incentives}, author = {Levanon, Sagi and Rosenfeld, Nir}, booktitle = {Proceedings of the 39th International Conference on Machine Learning}, pages = {12593--12618}, year = {2022}, editor = {Chaudhuri, Kamalika and Jegelka, Stefanie and Song, Le and Szepesvari, Csaba and Niu, Gang and Sabato, Sivan}, volume = {162}, series = {Proceedings of Machine Learning Research}, month = {17--23 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v162/levanon22a/levanon22a.pdf}, url = {https://proceedings.mlr.press/v162/levanon22a.html}, abstract = {Strategic classification studies learning in settings where self-interested users can strategically modify their features to obtain favorable predictive outcomes. A key working assumption, however, is that “favorable” always means “positive”; this may be appropriate in some applications (e.g., loan approval), but reduces to a fairly narrow view of what user interests can be. In this work we argue for a broader perspective on what accounts for strategic user behavior, and propose and study a flexible model of generalized strategic classification. Our generalized model subsumes most current models but includes other novel settings; among these, we identify and target one intriguing sub-class of problems in which the interests of users and the system are aligned. This setting reveals a surprising fact: that standard max-margin losses are ill-suited for strategic inputs. Returning to our fully generalized model, we propose a novel max-margin framework for strategic learning that is practical and effective, and which we analyze theoretically. We conclude with a set of experiments that empirically demonstrate the utility of our approach.} }
Endnote
%0 Conference Paper %T Generalized Strategic Classification and the Case of Aligned Incentives %A Sagi Levanon %A Nir Rosenfeld %B Proceedings of the 39th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2022 %E Kamalika Chaudhuri %E Stefanie Jegelka %E Le Song %E Csaba Szepesvari %E Gang Niu %E Sivan Sabato %F pmlr-v162-levanon22a %I PMLR %P 12593--12618 %U https://proceedings.mlr.press/v162/levanon22a.html %V 162 %X Strategic classification studies learning in settings where self-interested users can strategically modify their features to obtain favorable predictive outcomes. A key working assumption, however, is that “favorable” always means “positive”; this may be appropriate in some applications (e.g., loan approval), but reduces to a fairly narrow view of what user interests can be. In this work we argue for a broader perspective on what accounts for strategic user behavior, and propose and study a flexible model of generalized strategic classification. Our generalized model subsumes most current models but includes other novel settings; among these, we identify and target one intriguing sub-class of problems in which the interests of users and the system are aligned. This setting reveals a surprising fact: that standard max-margin losses are ill-suited for strategic inputs. Returning to our fully generalized model, we propose a novel max-margin framework for strategic learning that is practical and effective, and which we analyze theoretically. We conclude with a set of experiments that empirically demonstrate the utility of our approach.
APA
Levanon, S. & Rosenfeld, N.. (2022). Generalized Strategic Classification and the Case of Aligned Incentives. Proceedings of the 39th International Conference on Machine Learning, in Proceedings of Machine Learning Research 162:12593-12618 Available from https://proceedings.mlr.press/v162/levanon22a.html.

Related Material