Seminars and Conferences
Fall 2026 Colloquium Talks
Date: Thursday, November 5
Time: 1:10鈥�2:00 p.m.
Location: MSB 114
Speaker: Ethan AckelsbergTitle: Structure and randomness in combinatorics and dynamics
Abstract: Analytic and dynamical tools from ergodic theory are surprisingly adept at treating discrete, seemingly static problems in combinatorics and number theory. We will describe the deep connection between combinatorial patterns in subsets of the integers and recurrence phenomena in ergodic theory, as realized by Furstenberg in the 1970s. We will then explore how ideas from ergodic theory鈥攊n particular, the pervasive dichotomy between structure and randomness in dynamical systems鈥攑rovide deeper insight into combinatorial and number-theoretic problems and, conversely, how combinatorics and number theory enrich our understanding of the long-term statistical behavior of dynamical systems.Date: Thursday, October 22
Time: 1:10鈥�2:00 p.m.
Location: MSB 114
Speaker: Morley DavidsonTitle: "What is a Lean certification of a proof? And is it reliable?"
Abstract: Abstract: In this talk we'll examine the ever-increasing use of proof assistant software nowadays, particularly the role of Lean 4 in OpenAI's Oct. 6 posting of 700+ preprints with purported major breakthroughs including the quasi-Riemann hypothesis. To illustrate the pros and cons of automated proof certification, we'll focus on very recent AI+Lean experiments on irrationality of "odd zeta values," including an application of the new idea behind Fauzan's Sept.17 posting that zeta(5) is irrational, the first such result since Apery's 1978 proof for zeta(3), to showing (ostensibly!) that at least one of zeta(7), zeta(9), zeta(11), or zeta(13) is irrational. If its Lean certification holds water, this will update the 2001 breakthrough of Zudilin on four consecutive odd zeta values starting with zeta(5). But how can we be sure that a Lean certification is itself valid?
Date: Thursday, October 15
Time: 1:10鈥�2:00 p.m.
Location: MSB 114
Speaker: Fedor NazarovTitle: "What is a Lean certification of a proof? And is it reliable?"
Abstract: The formal result we will discuss is the following: Let T be a finite tree with the set V of vertices and E of edges. Consider the space 鈩�V of functions x: V 鈫� 鈩� with the norm鈥�x鈥� = 鈭� av|x(v)| + 鈭� ce触螖ex| v 鈭� V e 鈭� E where av, ce > 0 are arbitrary and 螖ex = |x(v鈥�) 鈭� x(v鈥�)| for e = {v鈥�, v鈥硙. Then this space is isomorphic to the usual 鈩�1 with the underlying Banach鈥揗azur constant at most 14. Recall that the B-M constant for two Banach spaces X and Y is justinf { 鈥�T鈥� 鈥�T鈭�1鈥� : T 鈭� 鈩�(X, Y), T invertible }.Some meta-mathematical discussion will be present too, so bring your brains! This is joint work with Micha艂 Wojciechowski.
Date: Thursday, October 1
Time: 1:10鈥�2:00 p.m.
Location: MSB 114
Speaker: Fedor NazarovTitle: "Very simple questions that can get complicated quickly鈥�.
Abstract: We'll discuss an elementary discrete geometry question that I encountered first in my 8th grade but have never resolved to my complete satisfaction since then. The talk will be interactive, so, please, bring your brains (in full compliance with our great non-discrimination laws, free and unrestricted usage of prosthetics, a.k.a. AI, will also be allowed, regardless of the manufacturer, the model, and the monthly rent rate, and would not be discriminated against in any way during the discussion).
Seminars
- Algebra Seminar
- Non-Linear Analysis Seminar
- Number Theory Seminar
- Computational and Applied Math Seminar
- Statistics Seminar
Past Conferences
Spring 2026
Informal Analysis Seminar August 2023
November 9-10, 2019
April 20-21, 2019
Feb. 9-10, 2019, 91PORNA
October 11-13, 2018
February 24-25, 2018
April 7-9, 2017October 28-30, 2016
April 30- May 1, 2016
November 14-15, 2015