We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: An Algorithmic Framework for Bias Bounties

Abstract: We propose and analyze an algorithmic framework for "bias bounties": events in which external participants are invited to propose improvements to a trained model, akin to bug bounty events in software and security. Our framework allows participants to submit arbitrary subgroup improvements, which are then algorithmically incorporated into an updated model. Our algorithm has the property that there is no tension between overall and subgroup accuracies, nor between different subgroup accuracies, and it enjoys provable convergence to either the Bayes optimal model or a state in which no further improvements can be found by the participants. We provide formal analyses of our framework, experimental evaluation, and findings from a preliminary bias bounty event.
Subjects: Machine Learning (cs.LG); Computers and Society (cs.CY); Data Structures and Algorithms (cs.DS)
Cite as: arXiv:2201.10408 [cs.LG]
  (or arXiv:2201.10408v4 [cs.LG] for this version)

Submission history

From: Ira Globus-Harris [view email]
[v1] Tue, 25 Jan 2022 15:57:48 GMT (2681kb,D)
[v2] Wed, 26 Jan 2022 03:58:00 GMT (2682kb,D)
[v3] Mon, 11 Apr 2022 19:37:46 GMT (2852kb,D)
[v4] Mon, 9 May 2022 19:44:06 GMT (2163kb,D)

Link back to: arXiv, form interface, contact.