Student Seminars: Yating Liu
10:00–11:00 am Jones 111
Thursday, May 1, 2025, at 10:00 AM, in Jones 111, 5747 S. Ellis Avenue
PhD Dissertation Proposal Presentation
Yating Liu, Department of Statistics, The University of Chicago
“High-Dimensional Inference Through Latent Structure”
Bahadur Memorial Lectures: John Lafferty (Day 2)
3:30–4:30 pm Jones 303
Title: Abstraction in Artificial and Natural Intelligence: Part II: Models, Mechanisms, and Experiments
Abstract: Reasoning in terms of relations, analogies, and abstraction is a hallmark of human intelligence. How can abstract symbols emerge from distributed, neural representations? One general approach uses an inductive bias for learning called the “relational bottleneck” that is motivated from principles of cognitive neuroscience. We present a framework that builds this inductive bias into machine learning models that transform distributed symbols to implement a form of abstraction. Computational experiments are presented on a broad range of problems. Biologically plausible mechanisms for these models are proposed to shed light on how abstraction may be implemented in the human brain.
Statistics Colloquium: Jiaying Gu
11:30 am–12:30 pm Jones 303
Jiaying Gu, Department of Economics, University of Toronto
Title: Empirical Bayes for Compound Adaptive Experiments
Abstract:
We investigate Empirical Bayes analysis in the context of compound adaptive experiments, where the arm distribution in each experiment follows a normal distribution with an unknown mean parameter that we aim to estimate. There are two primary approaches to EB estimation: g-modeling, which estimates the prior by maximizing the marginal likelihood, and f-modeling, which directly computes posterior means from the sample distribution of observations. We establish that g-modeling remains a valid EB procedure even when it incorrectly assumes exogenous data collection; it holds regardless of the sampling algorithm used and the endogeneity of sample sizes. One can apply standard g-modeling methods by treating the data as if it were exogenously sampled, and restrict attention to only the sample mean of the data. Remarkably, we show that the risk guarantees established for g-modeling with i.i.d data remain valid for adaptively generated data, with no need for prior knowledge of the sampling algorithm, even when it varies across experiments. In contrast, the f-modeling approach results in biased estimates. We validate the robustness of the g-modeling approach through simulations involving commonly used adaptive algorithms and illustrate its applicability using a real-world dataset comprising multiple sequential experiments.
Billingsley Lectures on Probability: Nina Holden
3:30–5:00 pm Jones 303
We are pleased to have Prof. Nina Holden, Courant Institute of Mathematical Sciences, New York University, as our honored speaker on Thursday, May 8, 2025, at 3:30 PM, Jones 303, 5747 South Ellis Avenue. A reception will immediately follow the lecture at 5:20 pm, in the Stevanovich Center for Financial Mathematics, 5727 South University Avenue.
Title: Scaling limits of random planar maps
Abstract: Planar maps are graphs embedded in the sphere such that no two edges cross, where we view two planar maps as equivalent if we can get one from the other via a continuous deformation of the sphere. In this talk we will present scaling limit results (i.e., convergence results) for random planar maps and we will focus in particular on a notion of convergence known as convergence under conformal embedding.
Statistics Colloquium: Anne van Delft
11:30 am–12:30 pm Jones 303
Anne van Delft, Department of Statistics, Columbia University
“TBA”
Statistics Colloquium: Michael Hudgens
11:30 am–12:30 pm Jones 303
Michael Hudgens, Department of Biostatistics, University of North Carolina at Chapel Hill
“Causal Inference in Infectious Disease Prevention Studies”