Course Schedule

• An introduction to the course; course logistics, and overview of the field of social choice and mechanism design.

• This topics provides an introduction to social choice theory and voting rules for preference data aggregation. We discuss the non-existence of "prefect" voting rules (Arrow's impossibility theorem) and the non-existence of intuitive but desirable rules that are immune to strategic manipulation (Gibbard-Satterthwaite theorem). A brief discussion of domain restriction (median rule) and complexity of manipulation (Bartholdi, Tovey and Trick condition).

Additional Material

• Economic reasoning and artificial intelligence [link]

• Representing and Reasoning with Preferences [link]

How should we allocate resources or goods when they are not divisible? We discuss a variety of deterministic relaxations such as envy-freeness up to one item as well as other fairness concepts such as the Maximin Share Guarantee, discuss computational challenges in achieving these properties, and explore their relation to other desirable concepts such as optimality (e.g. Pareto optimality, Nash welfare).

Additional Material

• To Divide the Rent, Start With a Triangle [link]

Can we achieve fairness through randomization? We discuss randomized algorithms for allocating indivisible resources and focus on two seminal mechanisms: Random Serial Dictatorship and the Probabilistic Serial Rule. We also discuss the relation to other deterministic fairness notions such as EF1.

Additional Material

• MatchU.ai [link]

• A New Solution to the Random Assignment Problem [link]

• Matching under Preferences [link]

Discussions on AI ethics, philosophical aspects of fairness and algorithmic decision making.

Additional Material

• A Survey on Bias and Fairness in Machine Learning (link)

• Fairness in Machine Learning: Lessons from Political Philosophy (link)

• Escaping the Impossibility of Fairness: From Formal to Substantive Algorithmic Fairness (link)

• October 21, 2026

  • Paper 1: Nash Learning from Human Feedback - Xinran / Patrice / Jenn

  • Project Disucssions

Additional Material

• PageRank: A trillion dollar algorithm (video) [link]

• November 4, 2026 

  • Paper 4: A Little Charity Guarantees Almost Envy-Freeness - Umar / Xiyuan

  • Paper 5: Judging LLM-as-a-judge with MT-bench and Chatbot Arena - Charles / Hins

• November 18, 2026 

  • Paper 7: Colosseum: Auditing Collusion in Cooperative Multi-Agent Systems -Sourajit / Christopher

  • Paper 8: Manipulating Gale-Shapley Algorithm: Preserving Stability and Remaining Inconspicuous - William / Xinnian / Rui

• December 2, 2026

  • Project Discussions

  • Project Presentations