Events

Joint Statistics and DSI Colloquium: Jingfeng Wu

4:00–5:00 pm DSI 105

Jingfeng Wu
Postdoctoral Fellow
University of California, Berkeley

Title:
Towards a Less Conservative Theory of Machine Learning: Unstable Optimization and Implicit Regularization

Abstract:
Deep learning’s empirical success challenges the “conservative” nature of classical optimization and statistical learning theories. Classical theory mandates small stepsizes for training stability and explicit regularization for complexity control. Yet, deep learning leverages mechanisms that thrive beyond these traditional boundaries. In this talk, I present a research program dedicated to building a less conservative theoretical foundation by demystifying two such mechanisms: 

1. Unstable Optimization: I show that large stepsizes, despite causing local oscillations, accelerate the global convergence of gradient descent (GD) in overparameterized logistic regression. 
2. Implicit Regularization: I show that the implicit regularization of early-stopped GD statistically dominates explicit $\ell_2$-regularization across all linear regression problem instances.

I further showcase how the theoretical principles lead to practice-relevant algorithmic designs (such as Seesaw for reducing serial steps in large language model pretraining). I conclude by outlining a path towards a rigorous understanding of modern learning paradigms.

Department of Computer Science and Data Science Institute Presents: Weijia Shi

2:30–3:30 pm DSI 105

Weijia Shi
PhD Candidate
University of Washington

Title: Breaking the Language Model Monolith

Abstract: Language models (LMs) are typically monolithic: a single model storing all knowledge and serving every use case. This design presents significant challenges; they often generate factually incorrect statements, require costly retraining to add or remove information, and face serious privacy and copyright issues. In this talk, I will discuss how to break this monolith by introducing modular architectures and training algorithms that separate capabilities across composable components. I’ll cover two forms of modularity: (1) External modularity, which augments LMs with external tools like retrievers to improve factuality and reasoning; and (2) internal modularity, which builds inherently modular LMs from decentrally trained components to enable flexible composition and an unprecedented level of control.

Student Seminar: Zixuan Qin

1:00–1:30 pm Jones 111

Wednesday, February 18, 2026, at 1:00 PM, in Jones 111, 5747 S. Ellis Avenue
Master’s Thesis Presentation
Zixuan Qin Department of Statistics, The University of Chicago
“Operator Learning and Bispectrum-Guided Diffusion for Functional Multi-Reference Alignment”

Student Seminar: Buning (Erica) Fan

1:30–2:00 pm Jones 111

Wednesday, February 18, 2026, at 1:30 PM, in Jones 111, 5747 S. Ellis Avenue
Master’s Thesis Presentation
Buning Fan, Department of Statistics, The University of Chicago
“Comparing Bayesian Software Platforms for Three-Level Mixed Effects Location Scale Models”

Student Seminar: Yushuo Li

2:00–2:30 pm Jones 111

Wednesday, February 18, 2026, at 2:00 PM, in Jones 111, 5747 S. Ellis Avenue
Master’s Thesis Presentation
Yushuo Li, Department of Statistics, The University of Chicago
“TBA”