This dissertation explores the intersection of privacy, fairness, and causality within the realm of machine learning (ML) and data-driven decision-making. The dissertation's main contributions can be summarized as follows: (1) we investigate the applicability of statistical and causality-based fairness notions in diverse application domains, evalua...
Recommender systems have the potential to reduce information frictions on the labor market by leveraging past interactions between job seekers and recruiters. This thesis presents several contributions to the design and evaluation of job recommender systems, leveraging detailed real-world data provided by the French Public Employment Service. First...
This thesis explores the importance of strengthening compliance with regulatory frameworks like the European General Data Protection Regulation (GDPR) in relation to data privacy and fairness in the field of voice biometrics focusing specifically on Automatic Speaker Verification (ASV). Through the use of cryptographic techniques, data perturbation...
Les pratiques alternatives de notation comme levier vers une évaluation plus équitable et qui accompagne mieux l'apprentissage
International audience
Disruption in the European airspace has a significant impact on all airlines operating in the affected regions. It results in a given number of flight cancellations and delays. When multiple airlines within a same group are faced with this same problem, our hypothesis is the following: we assume that jointly optimizing their resources could lead to...
Le dopage est une pratique qui porte atteinte à la santé du sportif et à l'équité des compétitions sportives. Le retentissement de certains scandales a conduit les autorités à mettre en place un droit répressif, pénal et administratif, particulièrement sévère. L'ampleur de la répression du dopage pose toutefois des difficultés au regard des droits ...
L'inclusion des élèves à besoins éducative particuliers dans les établissements scolaires, y compris en éducation physique et sportive (EPS), vise à garantir un accès équitable à l'éducation. Les adaptations pédagogiques et environnementales mise en place par l’enseignant, favorisent la participation active des EBEP, favorisant ainsi l'apprentissag...
Algorithms are increasingly used across all layers of society, including in high-stake decision systems, impacting individuals, cohorts and society as a whole. This omnipresence of algorithms raises important ethical and social concerns, including in particular privacy and fairness. In this thesis, we study these two technical subjects under the le...
This dissertation presents four published articles in the field of data ethics that extend our knowledge of fairness in machine learning and advance the state of the art of privacy in data collection and transmission.This document encompasses: (1) a general study of the trade-off between equal opportunity and accuracy of machine learning classifier...
