I am affiliated with The Wharton School of the University of Pennsylvania. My research centers on the intersection of economics and computer science and includes work on optimization and incentives. I live in San Francisco.
In addition to my academic interests I actively engage as a data science practitioner. I am the Head of Data Science at AngelList and I also serve as the co-Chief Scientist of Correlation One, a talent solutions technology company.
Along with Eric Budish, I developed Course Match, the course allocation system used by the Wharton School to quickly and fairly assign their MBA students to courses. Course Match has been covered by the Financial Times and Bloomberg Businessweek. For more technical detail on the algorithms that power Course Match, please see our paper in Operations Research.
Course Match was recently spun out of Wharton as Cognomos, a software service that offers course assignment and planning tools to university registrars. Cognomos performs course assignments at top schools like Toronto Rotman and Dartmouth Tuck.
I advise a number of startups, including Augur and Homebase. I typically invest in companies through my partnership in Indicator.
My involvement with Augur has taught me a great deal about blockchain theory and practice. In the popular media I have written an editorial on smart contracts for the FT Alphaville blog, and an editorial on rental attacks for TechCrunch.
I completed my PhD in Computer Science from Carnegie Mellon in 2012. In my thesis work, I explored policies for risk-sensitive portfolio optimization and pricing. My thesis was advised by Tuomas Sandholm, and my other committee members were Geoff Gordon, Javier Peña, Dave Pennock, and Pete Kyle. At CMU, my research was supported by the Google PhD Fellowship. My research mentor at Google was
Aranyak Mehta.
I have a degree in Applied Math from Harvard, where I worked with David Parkes and Uli Doraszelski.
Budish, E., Cachon, G., Kessler, J., and Othman, A. 2016. Course Match: A Large-Scale Implementation of Approximate Competitive Equilibrium from Equal Incomes for Combinatorial Allocation, in Operations Research (OR).
Combinatorial allocation involves assigning bundles of items to agents when the use of money is not allowed. Course allocation is one common application of combinatorial allocation, in which the bundles are schedules of courses and the assignees are students. Existing mechanisms used in practice have been shown to have serious flaws, which lead to allocations that are inefficient, unfair, or both. A new mechanism proposed by Budish [2011] is attractive in theory, but has several features that limit its feasibility for practice: reporting complexity, computational complexity, and approximations that can lead to violations of capacity constraints. This paper reports on the design and implementation of a new course allocation mechanism, Course Match, that enhances the Budish [2011] mechanism in various ways to make it suitable for practice. To find allocations, Course Match performs a massive parallel heuristic search that solves billions of Mixed-Integer Programs to output an approximate competitive equilibrium in a fake-money economy for courses. Quantitative summary statistics for two semesters of full-scale use at a large business school (Wharton, which has about 1,700 students and up to 350 courses in each semester) demonstrate that Course Match is both fair and efficient, a finding reinforced by student surveys showing large gains in satisfaction and perceived fairness.
Othman, A., Papadimitriou, C., and Rubinstein, A. 2016. The Complexity of Fairness Through Equilibrium, in ACM Transactions on Economics and Computation (TEAC).
Competitive equilibrium from equal incomes (CEEI) is a well-known fair allocation mechanism with desirable fairness and efficiency properties; however, with indivisible resources, a CEEI may not exist [Foley 1967; Varian 1974; Thomson and Varian 1985]. It was shown in Budish [2011] that in the case of indivisible resources, there is always an allocation, called A-CEEI, that is approximately fair, approximately truthful, and approximately efficient for some favorable approximation parameters. A heuristic search that attempts to find this approximation is used in practice to assign business school students to courses. In this article, we show that finding the A-CEEI allocation guaranteed to exist by Budish’s theorem is PPAD-complete. We further show that finding an approximate equilibrium with better approximation guarantees is even harder: NP-complete.
Othman, A., Pennock, D., Reeves, D., and Sandholm, T. 2013. A Practical Liquidity-Sensitive Automated Market Maker, in ACM Transactions on Economics and Computation (TEAC).
Automated market makers are algorithmic agents that enable participation and information elicitation in electronic markets. They have been widely and successfully applied in artificial-money settings, like some Internet prediction markets. Automated market makers from the literature suffer from two problems that contribute to their impracticality and impair their use beyond artificial-money settings: first, they are unable to adapt to liquidity, so that trades cause prices to move the same amount in both heavily- and lightly-traded markets, and second, in typical circumstances, they run at a deficit. In this paper, we construct a market maker that is both sensitive to liquidity and can run at a profit. Our market maker has bounded loss for any initial level of liquidity and, as the initial level of liquidity approaches zero, worst-case loss approaches zero. For any level of initial liquidity we can establish a boundary in market state space such that, if the market terminates within that boundary, the market maker books a profit regardless of the realized outcome.
Othman, A. and Sandholm, T. 2012. Inventory-based versus Prior-based Options Trading Agents, in Algorithmic Finance.
Options are a basic, widely-traded form of financial derivative that offer payouts based on the future price of an underlying asset. The finance literature gives us option-trading algorithms that take into consideration information about how prices move over time but do not explicitly involve the trades the agent made in the past. In contrast, the prediction market literature gives us automated market-making agents (like the popular LMSR) that are event-independent and price trades based only on the inventories the agent holds. We simulate the performance of five trading agents inspired by these literatures on a large database of recent historical option prices. We find that a combination of the two approaches produced the best results in our experiments: a trading agent that keeps track of previously-made trades combined with a good prior distribution on how prices move over time. The experimental success of this synthesized trader has implications for agent design in both financial and prediction markets.
