Each talk will be 20 minutes with 10 minutes for discussion. For each paper, we have invited experts on the topic to the talk to ensure a lively discussion.
“Learning Opinions in Social Networks” by Vincent Conitzer, Debmalya Panigrahi and Hanrui Zhang; Invited Discussant: Wei Chen and David Kempe; Chair: Grant Schoenebeck
Abstract: We study the problem of learning opinions in social networks. The learner observes the states of some sample nodes from a social network, and tries to infer the states of other nodes, based on the structure of the network. We show that sample-efficient learning is impossible when the network exhibits strong noise, and give a polynomial-time algorithm for the problem with nearly optimal sample complexity when the network is sufficiently stable.
“Towards Data Auctions With Externalities” by Anish Agarwal, Munther Dahleh, Thibaut Horel and Maryann Rui; Invited Discussants: Dirk Bergemann and Tan Gan; Chair: Heinrick Nax
Abstract: The design of data markets has gained in importance as firms increasingly use predictions from machine learning models to make their operations more effective, yet need to externally acquire the necessary training data to fit such models. This is particularly true in the context of the Internet where an ever-increasing amount of user data is being collected and exchanged. A property of such markets that has been given limited consideration thus far is the externality faced by a firm when data is allocated to other, competing firms. Addressing this is likely necessary for progress towards the practical implementation of such markets. In this work, we consider the case with n competing firms and a monopolistic data seller. We demonstrate that modeling the utility of firms solely through the increase in prediction accuracy experienced reduces the complex, combinatorial problem of allocating and pricing multiple data sets to an auction of a single digital (freely replicable) good. Crucially, this also enables us to model the negative externalities experienced by a firm resulting from other firms’ allocations as a weighted directed graph. We obtain forms of the welfare-maximizing and revenue-maximizing auctions for such settings. and highlight how the form of the firms’ private information – whether they know the externalities they exert on others or that others exert on them – affects the structure of the optimal mechanisms. We find that in all cases, the optimal allocation rules turn out to be single thresholds (one per firm), in which the seller allocates all information or none of it to a firm.
“A Closed-Loop Framework for Inference, Prediction and Control of SIR Epidemics on Networks” by Ashish R. Hota, Jaydeep Godbole, Sanket Kumar Singh and Philip E. Pare; Invited Discussants: Kuang Xu and Lei Ying; Chair: Longbo Huang
Abstract: Motivated by the ongoing pandemic COVID-19, we propose a closed-loop framework that combines inference from testing data, learning the parameters of the dynamics and optimal resource allocation for controlling the spread of the susceptible-infected-recovered (SIR) epidemic on networks. Our framework incorporates several key factors present in testing data, such as high risk individuals are more likely to undergo testing and infected individuals can remain as asymptomatic carriers of the disease. We then present two tractable optimization problems to evaluate the trade-off between controlling the growth-rate of the epidemic and the cost of non-pharmaceutical interventions (NPIs). Our results provide critical insights for policy-makers, including the emergence of a second wave of infections if NPIs are prematurely withdrawn.