Fairness and Alignment in Collective Decision-Making under Uncertainty
Abstract: Collective decision-making (CDM) is a process where multiple individuals often with diverse goals and preferences interact to arrive at a shared decision, like friends choosing a restaurant or political voting. These diverse and sometimes conflicting goals and preferences introduce complexities of alignment in the optimisation process. While current techniques attempt to tackle these issues it is unclear how online machine-learning approaches can be effectively integrated to optimize the CDM process in the absence of explicit feedback. In addition to this, it is crucial to make the collective decision fair for social cohesion and harmony. To address this challenge, we explore aggregation and preference elicitation methods that are fair and align collective decisions with the values of participants, emphasizing the importance of managing inherent uncertainties in both the outcomes of decisions and the preferences of decision-makers.
Short bio: Apurva Shah is a first-year Joint PhD student at Université Libre de Bruxelles and Vrije Universiteit Brussel in Belgium. She is currently working on fairness and alignment in collective decision-making under uncertainty. She completed her MSc in Behavioural and Data Science from the University of Warwick, UK and her BA in Psychology from Fergusson College, India.
Presenter: Apurva Shah
Date: 2025-02-28 12:30 (CET)
Location: Oficinas ELLIS Alicante, Muelle Pte., 5 – Edificio A, Alicante 03001, Alicante ES
This talk is part of the ELLIS / ULB-MLG Workshop.