Proportionality in Thumbs Up and Down Voting

Consider the decision-making setting where agents elect a panel by expressing both positive and negative preferences. Prominently, in constitutional AI, citizens democratically select a slate of ethical preferences on which a foundation model is to be trained. There, in practice, agents may both approve and disapprove of different ethical principles. Proportionality has been well-studied in computational social choice for approval ballots, but its meaning remains unclear when negative sentiments are also considered. In this work, we propose two conceptually distinct approaches to interpret proportionality in the presence of up and down votes. The first approach treats the satisfaction from electing candidates and the impact of vetoing them as comparable, leading to combined proportionality guarantees. The second approach considers veto power separately, introducing guarantees distinct from traditional proportionality. We formalize axioms for each perspective and examine their satisfiability by suitable adaptations of Phragm\'en's rule, Proportional Approval Voting rule and the Method of Equal Shares.

Core-Stable Committees under Restricted Domains

We study the setting of committee elections, where a group of individuals needs to collectively select a given size subset of available objects. This model is relevant for a number of real-life scenarios including political elections, participatory budgeting, and facility-location. We focus on the core -- the classic notion of proportionality, stability and fairness. We show that for a number of restricted domains including voter-interval, candidate-interval, single-peaked, and single-crossing preferences the core is non-empty and can be found in polynomial time. We show that the core might be empty for strict top-monotonic preferences, yet we introduce a relaxation of this class, which guarantees non-emptiness of the core. Our algorithms work both in the randomized and discrete models. We also show that the classic known proportional rules do not return committees from the core even for the most restrictive domains among those we consider (in particular for 1D-Euclidean preferences). We additionally prove a number of structural results that give better insights into the nature of some of the restricted domains, and which in particular give a better intuitive understanding of the class of top-monotonic preferences.