LU-UMN Joint Probability Seminar (Fall 2026)

I am co-organizing a virtual Probability Seminar with Wei-Kuo Chen and Arnab Sen at University of Minnesota for the Fall 2026 semester. Please email me to get the Zoom link for the seminar series.

The seminar is held on Fridays at 2:30pm (Eastern Time), unless otherwise noted below.

09/11/2026	First Week at UMN
09/25/2026	Byron Chin (Georgia Tech)
	Title: Solvability of the random stable roommates problem.
	Abstract: Assume that n = 2k potential roommates each have an ordered preference of the (n-1) others. A stable matching is a perfect matching of the n roommates in which no two unmatched people prefer each other to their matched partners. When the preferences are chosen uniformly at random, Gusfield and Irving predicted in 1989 that there is no stable matching with high probability. I will discuss our proof of this conjecture, based on a new approach to understanding the landscape of stable matchings. Joint work with Marcus Michelen.
10/02/2026	Fu-Husan Ho (Weizmann Institute)
	Title: Hopf formula for multispecies spin glasses.
	Abstract: The limiting free energy of multispecies spin glasses with nonconvex covariance function has been a challenge for mathematicians until recently. In this talk, I will briefly explain the theoretical framework to tackle this problem, and how it is reduced to understanding the concavity of a Parisi PDE. The talk will mostly base on my recent work and the concurrent work of HB Chen, Issa and Mourrat.
10/09/2026	Jeanne Boursier (MIT)
	Title: Large deviations for the TAP landscape of Ising spin glasses.
	Abstract: To understand spin glasses, physicists in the 1970s and 1980s began studying the landscape of a high-dimensional random function called the Thouless–Anderson–Palmer (TAP) free energy. Calculations imposing supersymmetry produced an elegant formula for counting critical points, but its probabilistic meaning remained unclear. I will explain how this formula governs rare events in the landscape: the probability that the maximum exceeds its typical value. The approach also provides a route to a constructive proof of Parisi’s formula for the limiting free energy of Ising spin glasses, extending ideas of Huang and Sellke for spherical models.
10/16/2026	Bhargav Narayanan (Rutgers)
	Title: Thresholds on the torus.
	Abstract: Sample n points uniformly from the two-dimensional torus and join points when they are at most some distance r apart. Does every graph property have a threshold in r, in the same way that every graph property has one in the edge density p as in G(n,p)? It turns out yet, yes, and in this talk, I will take the scenic route to the proof.
10/30/2026	Shuyang Gong (PKU/Yale)
	Title: Strong Freezing in the Binary Perceptron Models.
	Abstract: Binary perceptron models describe vertices of a high-dimensional hypercube satisfying random linear constraints. Statistical physics predicts a striking geometry: a typical solution is isolated from every other solution by a Hamming distance proportional to the dimension. In this talk, I will present a proof of the long-standing strong-freezing conjecture, first proposed by Krauth–Mézard (1989), later strengthened by Huang–Wong–Kabashima (2013). Our result settles the longstanding problem for the asymmetric binary perceptron at every margin throughout the satisfiable regime, while recovering the known symmetric case. More generally, the theorem applies to a broad class of perceptron models with independent disorder entries having mean zero, variance one. It provides stretched-exponential control of the exceptional fraction of solutions, with a similarly small probability of failure. The proof rests on a simple Poissonization idea that reveals an independence structure amenable to the FKG inequality, bypassing the need to control fluctuations in the number of solutions. This is joint work with Shuangping Li (Yale).
12/11/2026	Amanda Young (UIC)
	Title: TBA
	Abstract: TBA

←  homepage