Skip to content Skip to navigation

The RAIN seminar is held on Wednesdays from 12:00-1:00pm in Y2E2 101 . And yes, lunch is provided!

RAIN schedule for Winter Quarter 2017-18

Date Speaker Topic
Jan 17
Dean Eckles Randomization inference for spillovers in networks
Jan 31
Alex Peysakhovich Building a cooperator
Feb 14
Mohammad Akbarpours Diffusion, Seeding, and the Value of Network Information
Feb 28
Rachel Cummings TBA

Google Calendar for RAIN


Previous year's talks

Archived talks can be accessed here.

Talk Abstracts

Randomization inference for spillovers in networks
Dean Eckles, MIT

Abstract: Social and behavioral scientists are interested in testing of hypotheses about spillovers (i.e. interference, exogenous peer effects) in social networks; and similar questions may arise in other settings (e.g., biological and computer networks). However, when there is a single network, this is complicated by lack of independent observations. We explore Fisherian randomization inference as an approach to exact finite-population inference, where the main problem is that the relevant hypotheses are non-sharp null hypotheses. Fisherian randomization inference can be used to test these hypotheses either by (a) making the hypotheses sharp by assuming a model for direct effects or (b) conducting conditional randomization inference such that the hypotheses are sharp. I present both of these approaches, the latter of which is developed in Aronow (2012) and our paper (Athey, Eckles & Imbens, 2017). This usually involves selecting some vertices to be "focal" and conditioning on their treatment assignment and/or the assignment of some of all of their network neighbors. The selection of this set can present interesting algorithmic questions; we, for example, make use of greedy methods for finding maximal independent sets. I illustrate these methods with application to a large voter turnout experiment on Facebook (Jones et al., 2017).

Bio: Dean Eckles is a social scientist, statistician, and faculty at the Massachusetts Institute of Technology (MIT). Dean is the KDD Career Development Professor in Communications and Technology, an assistant professor in the MIT Sloan School of Management, and affiliated faculty at the MIT Institute for Data, Systems & Society. He was previously a member of the Core Data Science team at Facebook. Much of his work examines how interactive technologies affect human behavior by mediating, amplifying, and directing social influence - and statistical methods to study these processes. Dean's empirical work uses large field experiments and observational studies. His research appears in the Proceedings of the National Academy of Sciences and other peer-reviewed journals and proceedings in statistics, computer science, and marketing. Dean holds degrees from Stanford University in philosophy (BA), cognitive science (BS, MS), statistics (MS), and communication (PhD).


Building a cooperator
Alex Peysakhovich, Facebook

Abstract: Social dilemmas are situations where individuals face a temptation to increase their payoffs at a cost to total welfare. Building artificially intelligent agents that achieve good outcomes in these situations is important because many real world interactions include a tension between selfish interests and the welfare of others. We show how to modify modern reinforcement learning methods to construct agents that act in ways that are simple to understand, nice (begin by cooperating), provokable (try to avoid being exploited), and forgiving (try to return to mutual cooperation). We show both theoretically and experimentally that such agents can maintain cooperation in Markov social dilemmas. Our construction does not require training methods beyond a modification of self-play, thus if an environment is such that good strategies can be constructed in the zero-sum case (eg. Atari) then we can construct agents that solve social dilemmas in this environment.

Bio: Alex is a Research Scientist at Facebook Artificial Intelligence Research working on both human and machine decision-making. He got his PhD in Behavioral Economics from Harvard University and was a post-doc with David Rand at the Human Cooperation Lab. He spent several years working with the Facebook News Feed team on combining survey and behavioral data, natural language processing for detecting clickbait, and on building tools for advanced experimentation such as heterogeneous treatment effect detection.


Diffusion, Seeding, and the Value of Network Information
Mohammad Akbarpour, Stanford

Abstract: When communicating information to individuals is costly, policymakers try to identify the best 'seeds' to prompt a cascade of information within a social network. Numerous studies have proposed various network-centrality based heuristics to choose initial seeds in a way that is likely to boost diffusion. Here we show that, for a frequently studied diffusion process, randomly seeding s + x individuals can prompt a larger cascade than optimally seeding the best s individuals, for a small x. This suggests that the returns to collecting and analyzing network information to identify the optimal seeds may not be economically significant. Given these findings, practitioners interested in communicating a message to a large number of people may wish to compare the cost of identifying optimal seeds with the cost of informing a few additional people. Moreover, researchers studying network-based seeding heuristics may consider reporting 'extra seeds required by random seeding to achieve the same diffusion' as an easily interpretable information about the magnitude of their results.

Bio: Mohammad Akbarpour is an Assistant Professor of Economics at Stanford Graduate School of Business. His research bridges between economics and computer science, and is focused on market design, networked markets, and the economics of organ markets. Mohammad received his PhD from Stanford University Economics department in 2015 and his B.Sc in Electrical Engineering from Sharif University of Technology in Iran. As a consulting economist at Auctionomics, he has been involved in designing multiple spectrum auction markets around the globe. He has also been an instructor at Khan Academy Farsi, teaching hundreds of video-lectures in high school physics and calculus, and game theory.