To Pool or Not To Pool: Analyzing the Regularizing Effects of Group-Fair Training on Shared Models
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911786260758528 |
|---|---|
| author | Cousins, Cyrus Kumar, I. Elizabeth Venkatasubramanian, Suresh |
| author_facet | Cousins, Cyrus Kumar, I. Elizabeth Venkatasubramanian, Suresh |
| contents | In fair machine learning, one source of performance disparities between groups is over-fitting to groups with relatively few training samples. We derive group-specific bounds on the generalization error of welfare-centric fair machine learning that benefit from the larger sample size of the majority group. We do this by considering group-specific Rademacher averages over a restricted hypothesis class, which contains the family of models likely to perform well with respect to a fair learning objective (e.g., a power-mean). Our simulations demonstrate these bounds improve over a naive method, as expected by theory, with particularly significant improvement for smaller group sizes. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_18803 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | To Pool or Not To Pool: Analyzing the Regularizing Effects of Group-Fair Training on Shared Models Cousins, Cyrus Kumar, I. Elizabeth Venkatasubramanian, Suresh Machine Learning Computers and Society In fair machine learning, one source of performance disparities between groups is over-fitting to groups with relatively few training samples. We derive group-specific bounds on the generalization error of welfare-centric fair machine learning that benefit from the larger sample size of the majority group. We do this by considering group-specific Rademacher averages over a restricted hypothesis class, which contains the family of models likely to perform well with respect to a fair learning objective (e.g., a power-mean). Our simulations demonstrate these bounds improve over a naive method, as expected by theory, with particularly significant improvement for smaller group sizes. |
| title | To Pool or Not To Pool: Analyzing the Regularizing Effects of Group-Fair Training on Shared Models |
| topic | Machine Learning Computers and Society |
| url | https://arxiv.org/abs/2402.18803 |