Empirical Risk Minimization under Fairness Constraints
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14400
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 31 (NIPS 2018)
- First page
- 2796
- Volume
- 31
- Issue
- -
- ISSN
- 1049-5258
- Open access status
- Compliant
- Month of publication
- December
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
4
- Research group(s)
-
-
- Citation count
- 0
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Originality: This paper presents an efficient and principled approach to incorporate fairness requirements into a learning algorithm. Significance: There has been a lot of interest in algorithm fairness during the recent years. Previous work (featured in more than 500 research papers) has addressed almost exclusively heuristic approaches and practical applications. This is the first paper studying efficient classification methods which meet fairness requirements and is supported by prediction guarantees. Rigour: We present a trick to derive simultaneous fairness and error bounds for the method and provide a detailed experimental evaluation of its statistical performance, improving upon state-of-the-art approaches.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -