Computational Democracy Research Seminar - Eyal Briman, Ben Gurion University of Negev

Title: Application Oriented Social Choice with Structural Constraints

Abstracts:

Collective decision-making lies at the core of many consequential systems in modern society, including democratic elections, online governance, resource allocation, and increasingly, AI systems that interact with human groups. Computational Social Choice (COMSOC) provides rigorous tools for studying how individual preferences can be aggregated into collective decisions, and how to evaluate the fairness, efficiency, and computational tractability of these processes. Despite major advances, existing COMSOC models remain too restrictive for many real-world settings. Classical frameworks typically assume that the outcome is simple, such as a single winner, a fixed set, or a static allocation, and that the alternatives are explicitly given in advance. In practice, however, collective decisions often involve complex, structured outputs, such as collaborative documents, ranked lists, or multi-period allocations. Moreover, alternatives may themselves need to be generated, negotiated, or arranged during the decision process, and preferences may be expressed indirectly through delegation, proxy voting, or AI-assisted intermediaries. My research addresses these challenges by studying two interconnected forms of complexity: complex outputs and complex inputs. Addressing them requires new theoretical models, algorithmic techniques, and applications spanning collaborative content creation, democratic governance, fair resource allocation, and AI-mediated deliberation.