Robust ML Auditing using Prior Knowledge
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909620736360448 |
|---|---|
| author | Bourrée, Jade Garcia Godinot, Augustin De Vos, Martijn Vujasinovic, Milos Biswas, Sayan Tredan, Gilles Merrer, Erwan Le Kermarrec, Anne-Marie |
| author_facet | Bourrée, Jade Garcia Godinot, Augustin De Vos, Martijn Vujasinovic, Milos Biswas, Sayan Tredan, Gilles Merrer, Erwan Le Kermarrec, Anne-Marie |
| contents | Among the many technical challenges to enforcing AI regulations, one crucial yet underexplored problem is the risk of audit manipulation. This manipulation occurs when a platform deliberately alters its answers to a regulator to pass an audit without modifying its answers to other users. In this paper, we introduce a novel approach to manipulation-proof auditing by taking into account the auditor's prior knowledge of the task solved by the platform. We first demonstrate that regulators must not rely on public priors (e.g. a public dataset), as platforms could easily fool the auditor in such cases. We then formally establish the conditions under which an auditor can prevent audit manipulations using prior knowledge about the ground truth. Finally, our experiments with two standard datasets illustrate the maximum level of unfairness a platform can hide before being detected as malicious. Our formalization and generalization of manipulation-proof auditing with a prior opens up new research directions for more robust fairness audits. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_04796 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Robust ML Auditing using Prior Knowledge Bourrée, Jade Garcia Godinot, Augustin De Vos, Martijn Vujasinovic, Milos Biswas, Sayan Tredan, Gilles Merrer, Erwan Le Kermarrec, Anne-Marie Machine Learning Among the many technical challenges to enforcing AI regulations, one crucial yet underexplored problem is the risk of audit manipulation. This manipulation occurs when a platform deliberately alters its answers to a regulator to pass an audit without modifying its answers to other users. In this paper, we introduce a novel approach to manipulation-proof auditing by taking into account the auditor's prior knowledge of the task solved by the platform. We first demonstrate that regulators must not rely on public priors (e.g. a public dataset), as platforms could easily fool the auditor in such cases. We then formally establish the conditions under which an auditor can prevent audit manipulations using prior knowledge about the ground truth. Finally, our experiments with two standard datasets illustrate the maximum level of unfairness a platform can hide before being detected as malicious. Our formalization and generalization of manipulation-proof auditing with a prior opens up new research directions for more robust fairness audits. |
| title | Robust ML Auditing using Prior Knowledge |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2505.04796 |