« Discussion entre les mathématiques et la philosophie sur le thème de l’équité algorithmique »

Mode hybride: lien zoom.

For the next CRÉ-Obvia ethics and AI lunchtime conference, we welcome Véronique Tremblay, data scientist and responsible AI expert at Beneva and doctoral student in data science at HEC-Montréal.

The main principles of responsible AI all insist on the need to build equitable models. However, there is no consensus on the notion of fairness, either philosophically or mathematically. Since a perfectly and universally fair model is unattainable, how can we build models that respect the principle of fairness?

As a statistician and experienced data scientist, Véronique Tremblay will share with you the path that led her to realize the importance of philosophy in the search for an algorithm that could be described as fair, but also in the overall work of the data scientist. She will present (in French) the challenges encountered and some possible solutions, with the aim of opening up a dialogue between mathematics and philosophy.