Current Topics in Privacy Seminar: Amin Rahimian скачать в хорошем качестве

Current Topics in Privacy Seminar: Amin Rahimian

Amin Rahimian - University of Pittsburgh Title :Privacy-Aware Sequential Learning Sequential learning examines inferences of individuals when they are making decisions based on their private information and the past actions of others. In privacy-preserving sequential learning, agents add endogenous noise to their actions to hide private signals. We study how privacy constraints affect learning across signal structures. Efficient social learning relies on information flow, seemingly in conflict with privacy. Surprisingly, with continuous signals and a fixed privacy budget (ε), the optimal randomization strategy balances privacy and accuracy, accelerating sequential learning to θ_ε(log n), faster than the non-private θ(√log(n)) rate. In the nonprivate baseline, both the expected time to the first correct action and the number of incorrect actions are unbounded, whereas in the private case with sufficiently small ε, both are finite. Sequential learning with continuous signals becomes more efficient under privacy because, under the false state, agents are more likely to randomize their actions given a higher probability of receiving signals contradicting the majority; privacy randomization then asymmetrically amplifies the log-likelihood ratio between two states, enhancing information aggregation. In heterogeneous populations, an order-optimal θ(√n) rate is achievable when some agents have very low privacy budgets. With binary signals, privacy reduces informativeness and impairs learning compared to the nonprivate baseline, though the relationship with privacy budget is nonmonotone. Our results show how privacy reshapes information dynamics in sensitive decision domains and have implications for online platform design and privacy policy. Joint work with Yuxin Liu BIo Amin Rahimian has been an assistant professor of industrial engineering at the University of Pittsburgh since 2020, where he leads the sociotechnical systems research lab and is also affiliated with the Pitt Cyber Institute and the Intelligent Systems program, as well as the MIT Initiative on Digital Economy. Prior to joining Pitt, he was a postdoc with joint appointments at MIT Institute for Data, Systems, and Society (IDSS) and MIT Sloan School of Management. He received his PhD in Electrical and Systems Engineering from the University of Pennsylvania and a Master’s in Statistics from the Wharton School. Broadly speaking, his works are at the intersection of networks, data, and decision sciences, and have been published in the Proceedings of the National Academy of Sciences, Nature Human Behaviour, Nature Communications, and the Operations Research journal, among others. His research interests are in applied probability, applied statistics, algorithms, decision and game theory, with applications in online social networks, public health, and e-commerce.