Department of Mathematics and Statistics, Binghamton University seminars

Algebra Seminar

2026-05-13T23:47:10-04:00

The Algebra Seminar

The seminar will meet in-person on Tuesdays in room WH-100E at 2:45 p.m. There should be refreshments served at 3:45 in our new lounge/coffee room, WH-104. Masks are optional.

Anyone wishing to give a talk in the Algebra Seminar this semester is requested to contact the organizers at least one week ahead of time, to provide a title and abstract. If a speaker prefers to give a zoom talk, the organizers will need to be notified at least one week ahead of time, and a link will be posted on this page.

A new link controlled by the organizers will be needed for zoom meetings of the Algebra Seminar.

Algebra Seminar Zoom Meeting Link

Organizers: Daniel Studenmund and Hung Tong-Viet

To receive announcements of seminar talks by email, please email one of the organizers with your name, email address and reason for joining this list if you are external to Binghamton University.

Fall 2026

August 18
Organizational Meeting

Please think about giving a talk in the Algebra Seminar, or inviting an outside speaker.

August 25
(? University)
Title

Abstract: Text of Abstract

September 1
(? University)
Title

Abstract: Text of Abstract

September 8
No Meeting (Monday Classes Meet)

September 15
(? University)
Title

Abstract: Text of Abstract

September 22
(? University)
Title

Abstract: Text of Abstract

September 29
(? University)
Title

Abstract: Text of Abstract

October 6
(? University)
Title

Abstract: Text of Abstract

October 13
No Meeting (Fall Break)
Title

Abstract: Text of Abstract

October 20
(? University)
Title

Abstract: Text of Abstract

October 27
(? University)
Title

Abstract: Text of Abstract

November 3
(? University)
Title

Abstract: Text of Abstract

November 10
(? University)
Title

Abstract: Text of Abstract

November 17
(? University)
Title

Abstract: Text of Abstract

November 24
No Meeting (Friday Classes Meet)

December 1
(? University)
Title

Abstract: Text of Abstract

December 8
(? University)
Title

Abstract: Text of Abstract

The Analysis Seminar

2026-05-03T15:40:42-04:00

The Analysis Seminar

Fourier

The seminar meets Wednesdays in WH-100E at 4:00-5:00 p.m. There are refreshments and snacks in WH-102 at 3:15.

Organizers:
Faculy:Paul Loya, David Renfrew, Minghao Rostami, Emmett Wyman, Xiangjin Xu, Ziyao Xu and Gang Zhou
Post-Docs: Rohan Sarkar

Previous talks

Fall 2014 to Fall 2025

Spring 2026

* January 21st, Wednesday (4-5pm)

Speaker :
Topic: organizational meeting

Abstract:

* January 28th, Wednesday (4-5pm)

Speaker : Chad Nelson (Binghamton University)
Topic: Fredholmness of Elliptic Operators on Manifolds with Boundary

Abstract: The classical calculus of pseudodifferential operators extends differential operators in a way that is suited to the construction of parametrices (pseudo-inverses) for elliptic operators. A fundamental consequence is that elliptic operators are Fredholm between appropriate Sobolev spaces on compact manifolds.

On manifolds with boundary, this implication no longer holds. Melrose’s calculus of b-pseudodifferential operators is the analogous class of operators which leads to Fredholm properties for elliptic operators satisfying a certain condition related to the boundary. In this talk, I will compare the classical case and the boundary case, emphasizing the new features introduced by the boundary—most notably the b-stretched product and the indicial operator—and explain how these lead to Fredholmness on weighted b-Sobolev spaces.

* February 4th, Wednesday (4-5pm)

Speaker : Emmanuel Adara (Binghamton University)
Topic: On Methods of Solution to Chemical Master Equation in Biochemical Systems

Abstract: In chemical kinetics, accurately modeling the dynamic behavior of chemical systems is essential for predicting reaction outcomes and optimizing processes. However, the challenge known as the “curse of dimensionality” has posed significant difficulties for conventional techniques employed in addressing the chemical master equation (CME). This predicament arises when the state space of the Markov chain expands exponentially with the number of species, rendering the CME computation practically unsolvable.

In this talk, I will discuss some methods of solving the CME, including Gillespie’s algorithm, the Chemical Langevin Equation, and the Method of Moments, along with an overview of tensor train and machine learning-based methods, which offer promising strategies for gaining insights into complex biological systems.

* Tuesday, February 10 (Joint with Data Science Seminar) (12:15-1:15pm)

Speaker : Dr. Yizeng Li (Department of Biomedical Engineering at Binghamton University)
Topic: Multiphase Continuum Models for Cell Migration.

Abstract: Cell migration is a fundamental process in physiology and disease, yet it poses challenging problems in multiscale modeling and continuum mechanics. Cell motility arises from the coupling of intracellular transport, active force generation, and evolving geometry. Cytoskeletal dynamics, in particular actin turnover and force production, provides a rich setting for mathematical analysis. In this talk, I will present a mathematical framework for mammalian cell motility based on multiphase continuum models with moving boundaries. The formulation incorporates fluid-structure interaction and active stresses to describe the coupled evolution of cytoskeletal flow and cell shape. The model predicts how migration efficiency depends on actin dynamics and geometric features of the cell. If time permits, I will also present a mechanical-electrical-chemical coupled model for water-driven cell motility induced by polarized membrane ion transport. This second framework highlights how transport processes and force balance together generate directed motion.

Biography of the speaker: Yizeng Li is an Assistant Professor in the Department of Biomedical Engineering at Binghamton University. She received MS from Mathematics and PhD from the Department of Mechanical Engineering at the University of Michigan-Ann Arbor. Afterwards, she was a postdoctoral researcher at Johns Hopkins University's Department of Mechanical Engineering and Institute for NanoBioTechnology. Her backgrounds are in theoretical mechanics and applied mathematics with applications to biophysics and mechanobiology. Li develops physiology-based mathematical models for cell motility, polarization, volume regulation, electro-homeostasis, signal transduction, and other biophysics problems. She also combines mathematical models with experimental data to explain non-intuitive cell biology phenomena.

* (Updated) March 19th, Thursday (4-5pm)

Speaker: Zheng Sun (University of Alabama, Tuscaloosa)
Topic: On a numerical artifact of solving shallow water equations with a discontinuous bottom

Abstract: The nonlinear shallow water equations are used to model the free surface flow in rivers and coastal areas for which the horizontal length scale is much greater than the vertical length scale. They have wide applications in oceanic sciences and hydraulic engineering. In this talk, we study a numerical artifact of solving the shallow water equations over a discontinuous riverbed. For various first-order methods, we report that the numerical solution will form a spurious spike in the numerical momentum at the discontinuous point of the bottom. This artifact will cause the convergence to a wrong solution in many test cases. We present a convergence analysis to show that this numerical artifact is caused by the numerical viscosity imposed at the discontinuous point. Motivated by our analysis, we propose a numerical fix which works for the nontransonic problems.

* March 25th, Wednesday (4-5pm)

Speaker: Yiming Zhao(Syracuse University)
Topic: $SL(n)$-invariant isoperimetric and Minkowski problems

Abstract: In convex geometry, where convex bodies are the primary objects of study, a central goal is to discover geometric invariants and measures that can be used to recover or characterize their shapes. Two intertwined lines of research pursue this objective. Isoperimetric inequalities involving geometric invariants, including the classical isoperimetric inequality and the celebrated Brunn–Minkowski inequality, seek to identify special shapes as extremals. Minkowski problems, a family of problems originating in the work of Minkowski, aim to recover, sometimes uniquely, the shape of an arbitrary convex body by solving measure equations that, under additional but unnecessary smoothness assumptions, reduce to Monge–Ampère-type equations. In this talk, after giving some historical background, I will discuss recent joint work with Dongmeng Xi that continues this line of research through the study of integral affine surface area and radial mean bodies.

* April 1st, Wednesday (4-5pm)

Speaker: Spring Break (Binghamton University)
Topic: A lower bound on sleep duration under optimal conditions

Abstract:

We show that in the absence of homework, sleep duration increases without bound. Applications to stress reduction are discussed.

* April 8th, Wednesday (4-5pm)

Speaker: Gang Zhou (Binghamton University)
Topic: Exact characterizations for quantum conditional mutual information and some other entropies

Abstract: I will present my latest results on quantum information theory. I will start with an introduction to the quantum theory for preparation, and then present the results.

Lieb and Ruskai's strong subadditivity theorem, which shows that the conditional mutual information must be nonnegative, is fundamental in quantum theory. It has numerous applications, such as in quantum error correction.

When the mutual information is zero, the Petz recovery map can be used to reconstruct the quantum channel.

When the mutual information is small, one seeks to define an optimal recovery channel. To this end, a mathematical characterization of the mutual information is desirable. In my latest paper I provided an exact characterization of the mutual information, along with characterizations for other entropies. My controls are sharp, leaving no room for improvement, in the sense that I provided equalities, regardless of whether the mutual information (or remainder) is small or large.

I transformed the definitions of these entropies into a summation of explicitly constructed terms, and the definition of each term obviously demonstrates the desired positivity/convexity/concavity. The summation converges rapidly and absolutely in a chosen elementary norm.

An exposition for the general public was provided by git.science and emailed to me, see the link https://gist.science/paper/2603.14650

* April 15th, Wednesday, 4:00-6:00pm (PhD Thesis Defense)

Speaker: Brian Kirby(Binghamton University)
Topic: On Resolving the Singularities of the Reissner-Nordström Penrose Diagram via Method of Blow-Ups

Abstract: The Penrose Diagram was first developed by Roger Penrose in the early 1960’s, taking an infinite spacetime and representing it in a concise diagram. This was first used to model the Schwartzchild spacetime metric, a black hole with no electric charge. The Penrose Diagram for the Reissner-Nordstrom metric was a result of the works of Brandon Carter, Stephen Hawking, and George Ellis, primarily in the late 1960’s to early 1970’s. These diagrams are still used today, primarily by physicists and astronomers to help better understand black holes, event horizons, and causality. This thesis focuses on better understanding the Penrose Diagram for the Reissner-Nordstrom metric. While work has been done recently to develop algorithms and explicitly construct these diagrams, current understanding of Penrose Diagrams works only on the interior of the diagrams, giving an incomplete picture of spacetime with infinitely many discontinuities with highly discontinuous behavior. In this thesis, we generalize the Reissner-Nordstrom metric and expand on its standard Penrose Diagram construction. Through the method of blow-ups of manifolds with corners, we resolve each singularity in the standard Penrose Diagram, and classify the asymptotic behavior of the metric at $r = 0$. In particular, we prove:

Theorem: Any Reissner-Nordstrom metric lifts to its blown-up Penrose Diagram to be a $C^{\infty}$ b-metric everywhere, up to and including the front faces, except at the $r = 0$ hypersurface, where it has an expansion of the form $r^{-2/3} F(r^{1/3})$.

The Thesis Committee members are: Paul Loya (Chair and Faculty Advisor), Emmett Wymann, Xiangjin Xu and Bruce White (Outside Examiner).

* April 22th, Wednesday (4-5pm)

Speaker: Brian Kirby(Binghamton University)
Topic: Explicit Construction of the Asymptotics of the Blown up Reissner Nordstrom Penrose Diagram

Abstract: In my thesis defense last week, I described the metric on the Blown up Reissner Nordstrom Penrose Diagram. In this talk, we fill in the details concerning the Asymptotics of the singularities.

April 29th, Wednesday (4-5pm)

Speaker: Yahong Yang (Georgia Institute of Technology)
Topic: Multiscale Neural Networks for Approximating Green’s Functions and Operators

Abstract: Neural networks (NNs) have been widely used to solve partial differential equations (PDEs) with broad applications in physics, biology, and engineering. One effective approach for solving PDEs with a fixed differential operator is to learn the associated Green’s function. However, Green’s functions are notoriously difficult to approximate due to their poor regularity, often requiring large neural networks and long training times.

In this talk, we address these challenges by leveraging multiscale neural networks to learn Green’s functions efficiently. Through theoretical analysis based on multiscale Barron space techniques, together with numerical experiments, we show that the multiscale approach significantly reduces the required network size and accelerates training. We then extend this framework to operator learning, enabling neural networks to efficiently and accurately learn the mapping from coefficient functions to Green’s functions.

* May 5th, Tuesday, 3:00-5:00pm (PhD Thesis Defense)

Location:Room 309 (Special Room)

Speaker: Paul Barber (Binghamton University)
Topic: On the dynamics of generic singularities formed by a semi-linear heat equation.

Abstract: We consider the generic type-I singularities produced by the one-dimensional nonlinear heat equation $\partial_t u= \partial_x^2 u+ u^3$. Assuming only that we have generic type-I blowup at a point and the initial data is positive, we are able to prove that this blowup point is isolated. Moreover, we are able to get a detailed description of the solution’s profile in a spacetime neighborhood of this point. This improves on previous well known results which only obtain estimates in a region of such a point that shrinks as one approaches the blowup time.

Our technique uses estimates on a propagated solution that initiates a bootstrap argument. Although we work with the specific equation above, we expect the techniques used will be applicable to other similar nonlinear parabolic equations, some of which occur in geometric flows.

The Thesis Committee consists of Gang Zhou (chair), Michael Dobbins, David Renfrew, Fake (Frank) Lu(outside examiner).

* May 6th, Wednesday (4-5pm) (Master Thesis Defense)

Speaker:Santiago Alzate (Binghamton University)
Topic: The Peter-Weyl's theorem and its applications

Abstract: This talk develops the foundations of harmonic analysis on compact Lie groups. After recalling necessary background in analysis on $\mathbb{R}^n$, topology, algebras and group theory, and the differential structure of Lie groups and their tangent spaces, we introduce Fourier analysis on $\mathbb{R}^n$ (Fourier transform, convolutions, distributions and Schwartz spaces) as preparatory material. Endowing a compact group with Haar measure, we present the Peter–Weyl theorem and use finite-dimensional unitary representations to construct Fourier series, trigonometric polynomials and the group Fourier transform. Using the Lie algebra structure we define the Casimir element (an analogue of the Laplacian) and show that Peter–Weyl components are its eigenspaces. The main analytic result is that the Fourier transform yields a surjective isometry between $L^2(G)$ and the corresponding $L^2$-space on the unitary dual, with an explicit inverse; we also study mapping properties on $C^\infty$, tempered distributions and Schwartz-type spaces on the dual. Finally, symbols and pseudo-differential operators on compact Lie groups are introduced and analyzed via the Fourier calculus developed here.

The examining committee is Laura Anderson, Paul Loya (Chair), and Xiangjin Xu.

The Arithmetic Seminar

2026-04-30T19:36:14-04:00

The Arithmetic Seminar

TOPICS: Arithmetic in the broadest sense that includes Number Theory (Elementary Arithmetic, Algebraic, Analytic, Combinatorial, etc.), Algebraic Geometry, Representation Theory, Lie Groups and Lie Algebras, Diophantine Geometry, Geometry of Numbers, Tropical Geometry, Arithmetic Dynamics, Arithmetic Topology, etc.

PLACE and TIME: This semester the seminar meets primarily on Tuesdays at 4:00 pm, with possible special lectures at other days and times. The in-house talks will be in-person, while visitors outside of Binghamton area will be in-person or by Zoom: Zoom link

ORGANIZERS:
Regular Faculy: Alexander Borisov, Marcin Mazur, Adrian Vasiu.
Post-Docs: Huy Dang

Current Ph.D. students: Hari Asokan, Mithun Padinhare Veettil.

Graduated Ph.D. students (in number theory and related topics): Ilir Snopce (Dec. 2009), Xiao Xiao (May 2011), Jinghao Li (May 2015), Ding Ding (Dec. 2015), Patrick Milano (May 2018), Changwei Zhou (May 2019), Patrick Carney (May 2023), Sarah Lamoureux (Sep. 2023), Sayak Sengupta (May 2024).

SEMINAR ANNOUNCEMENTS: To receive announcements of seminar talks by email, please join our mailing list.

Related seminar: Upstate New York Online Number Theory Colloquium (online, irregular): http://people.math.binghamton.edu/borisov/UpstateNYOnline/Colloquium.html

Previous Arithmetic Seminar Talks

————————- Fall 2025
Spring 2025 ——– Fall 2024
Spring 2024 ——– Fall 2023
Spring 2023 ——– Fall 2022
Spring 2022 ——– Fall 2021
Spring 2021 ——– Fall 2020
Spring 2020 ——– Fall 2019
Spring 2019 ——– Fall 2018
Spring 2018 ——– Fall 2017
Spring 2017 ——– Fall 2016
Spring 2016 ——– Fall 2015
Spring 2015 ——– Fall 2014

Spring 2026

January 27
Speaker: NA
Title: Organizational Meeting
Abstract:

February 3 (2:45-3:45 pm, cross-listed from Algebra Seminar)
Speaker: Tim Riley (Cornell University)
Title: Conjugator length
Abstract: The conjugacy problem for a finitely generated group $G$ asks for an algorithm which, on input a pair of words u and v, declares whether or not they represent conjugate elements of $G$. The conjugator length function $CL$ is its most direct quantification: $CL(n)$ is the minimal $N$ such that if $u$ and $v$ represent conjugate elements of $G$ and the sum of their lengths is at most $n$, then there is a word $w$ of length at most $N$ such that $uw=wv$ in $G$. I will talk about why this function is interesting and how it can behave, and I will highlight some open questions. En route I will talk about results variously with Martin Bridson, Conan Gillis, and Andrew Sale, as well as recent advances by Conan Gillis and Francis Wagner.

February 10
Speaker: Alexander Borisov (Binghamton)
Title: A structure sheaf for Kirch topology, an update
Abstract: Kirch topology on $\mathbb N$ goes back to a 1969 paper of Kirch. It can be defined by a basis of open sets that consists of all infinite arithmetic progressions $a+d\mathbb N_0$, such that $gcd(a,d)=1$ and $d$ is square-free. It is Hausdorff, connected, and locally connected. I will give an update on my current work on a natural presheaf of functions on this topological space: locally integer polynomial functions. In particular, I will discuss when the sheafification is equal to the presheaf, and when it is bigger. I will also discuss (Cech) cohomology. In particular, I will give examples with trivial and nontrivial H^1. No prior knowledge of the topic is assumed. This talk will also serve as an introduction to Mithun's talk next week.

February 17
Speaker: Mithun Veettil (Binghamton)
Title: Some results on the Locally LIP functions
Abstract: Locally LIP functions are obtained as a result of sheafification of the presheaf LIP on some infinite subset $X$ of $N={1,2,3,\ldots}$, with a prescribed topology. Often we work with Kirch topology on $N$ that makes $N$ a connected, locally connected, and Hausdorff space.
If the set $X$ is a union of non-connected open sets, then we can easily define a locally LIP function on $X$ that is not a LIP function globally. In fact, even if the space $X$ is connected, a locally LIP function on $X$ need not be a LIP function on $X$. In this talk, we will look at $X=N$\ $6N$, which is connected, and construct a locally LIP function that is not LIP on $X$. Also, we will show that this is not the case if one works with $\mathbb{Z}[1/2]$ instead of $\mathbb{Z}$ for the above set $X$.

February 24
Speaker: Hari Asokan (Binghamton)
Title: Variation of Geometric Invariant Theory
Abstract: Geometric Invariant Theory is used to construct quotients of group actions on varieties, but the outcome depends on a choice of linearization. Variation of Geometric Invariant Theory (VGIT) studies the different quotients resulting from changing this choice. In this talk I will give an informal introduction to VGIT, focusing on how stability changes as linearization varies.

March 3
Speaker: Connor Stewart (CUNY)
Title: Conductor–Discriminant Inequality for Tamely Ramified Cyclic Covers
Abstract: We consider $\mathbb{Z}/n$-covers $X\to\mathbb{P}^1$ defined over discretely valued fields $K$ with excellent valuation ring $\mathcal{O}_K$ and perfect residue field of characteristic not dividing $n$. Two standard measures of bad reduction for such a curve $X$ are the Artin conductor of its minimal regular model over $\mathcal{O}_K$ and the valuation of the discriminant of a Weierstrass equation for $X$. We prove an inequality relating these two measures. Specifically, if $X$ is given by an affine equation $y^n = f(x)$ with $f(x) \in \mathcal{O}_K[x]$, and if $\mathcal{X}$ is its minimal regular model over $\mathcal{O}_K$, then the negative of the Artin conductor of $\mathcal{X}$ is bounded above by $(n-1)v_K(\textrm{disc (rad}\ f))$. This extends previous work of Ogg, Saito, Liu, Srinivisan, and Obus-Srinivasan on elliptic and hyperelliptic curves. (Joint work with Andrew Obus and Padmavathi Srinivasan.)

March 10
Speaker: Eric Yin (Binghamton)
Title: Generating abelian extensions with elliptic curves with complex multiplication.
Abstract: An elliptic curve with complex multiplication is one with extra endomorphisms, ones that are not simply given by multiplication-by-m maps. In this talk we discuss how this extra structure allows us to find an analogy to Kronecker-Weber, generating abelian extensions of imaginary quadratic fields through torsion points on elliptic curves with CM. In addition, we discuss the role the ideal class group plays in both measuring ramification and classifying elliptic curves with a given endomorphism ring. This lets us describe both the Hilbert class field, the maximal unramified abelian extension, and the maximal abelian extension of any imaginary quadratic field.

March 17
Speaker: Anitha Srinivasan (Comillas University, Madrid), by Zoom
Title: The generalized Markoff equation
Abstract: The talk will look at various aspects of the generalized Markoff equation $a^2+b^2+c^2=3abc+m$ ($m\ge 0$), giving an overview of all the exciting work in the area. A few examples of topics that will be mentioned are: the classification of solution triples $(a, b, c)$ that come from $k$-Fibonacci sequences, open conjectures (which $m's$ have no solutions?), counting algorithms for the number of solutions (trees) and the Markoff equation mod $p$.

March 24
Speaker: Jauing Jun (SUNY New Paltz)
Title: Categorical approach to stability of tropical toric vector bundles
Abstract: Developing a suitable notion of vector bundles in tropical geometry has recently attracted considerable interest. In a recent work, Khan and Maclagan introduced tropical vector bundles using matroids, inspired by Klyachko’s classification of toric vector bundles, and studied their stability properties. In this talk, we reinterpret their notion of stability through the framework of the categorical approach to stability proposed by André. This perspective clarifies the structure underlying their results and places them in a broader conceptual setting. This is joint work with Alex Sistko and Cameron Wright.

April 14
Speaker: Anubhav Nanavaty (Cornell University)
Title: Feynman Integrals and Symmetric Matrices
Abstract: Since the foundational work of Broadhurst and Kreimer, there has been a significant push to understand the amplitudes of Feynman Integrals as periods, or integrals of algebraic functions over algebraic domains. Mysteriously, some of these amplitudes are special values of the Riemann zeta function (and more generally multiple zeta values). Work of Brown suggests that these amplitudes are related to the homology of Kontesevich's graph complex, and therefore the cohomology of the moduli space of curves by work of Chan, Galatius and Payne. Central to the story are the Borel classes - GLn(Z) equivariant cohomology classes on the space of projective symmetric matrices of full rank. I first show that, in a very general setting, the Voevodsky motive of this space splits into a direct sum of Tate Motives. I conclude by work in-progress with collaborators, where we aim to compute the weights of the Borel classes viewed as differential forms on the space of projective complex symmetric matrices of full rank and use this computation to express certain classes of Feynman amplitudes to multiple zeta values.

April 21
Speaker: Alexander Borisov (Binghamton)
Title: A structure sheaf for Kirch topology: cohomology theory
Abstract: This is an update on my work on locally LIP functions, as a sheaf for Kirch topology. After reviewing the definitions and some previous results, I will talk about most recent advances. In particular, I will prove that $H^1$ is zero on the basic open sets.

April 28
Speaker: Chaitanya Joglekar (Binghamton)
Title: Factoring polynomials using the LLL algorithm
Abstract: The LLL algorithm finds a “short enough” vector in a lattice in polynomial time. LLL has many applications, and among other things it can be used to factor polynomials with integer coefficients in polynomial time, by combining it with the Berlekamp algorithm for factoring polynomials over finite fields. In this talk I will cover the algorithm to factor polynomials, using LLL as a black box.

May 7 4:10-6:10 pm Special Event: PhD Defense
Speaker: Hari Asokan (Binghamton)
Title: Invariant Theory of Triplets of Quadratic Forms
Abstract: We study the invariant theory of triplets of quadratic forms over an algebraically closed field K under the congruence action of SL_n. This work is motivated by questions arising in delta-invariant theory for Hecke correspondences (see BV), which lead to the problem of comparing the full invariant K-algebra with the subalgebra generated by determinant polarizations (the Theta-invariants). This reduces to the analysis of the isospectral condition: the characteristic polynomial of D + tQ equals the characteristic polynomial of D for all t in K, where D is a diagonal matrix with distinct eigenvalues and Q is symmetric. We prove that if the characteristic of K is not 2, the only solution is Q = 0 if and only if n ≤ 4, while for n ≥ 5 nontrivial solutions exist. For each n ≥ 5, we construct explicit examples and establish lower and upper bounds on the dimension and rank of families of solutions; for instance, we show that for each D there exists a vector space M over K of dimension binomial(floor((n−1)/2), 2) consisting of symmetric matrices such that for each Q in M, the isospectral condition holds.
We also analyze the characteristic 2 case, where additional families of solutions appear. These results clarify the algebraic structure of invariants of quadratic forms and the limitations of determinant polarizations. In particular, the case n = 4 finds application in delta-geometry, as it ensures generic finiteness of the associated morphisms and enables explicit degree computations.

Colloquium

2017-01-05T18:20:20-04:00

Colloquium

Unless stated otherwise, colloquia are scheduled for Thursdays 4:30-5:30pm in WH-100E with refreshments served from 4:00-4:25 pm in WH-102.

Organizers: Thomas Farrell and Andrey Gogolev

Previous talks

Spring 2015

February 19, 4:30pm
Speaker: Jonathan Williams (University of Georgia)
Topic: A new approach to general smooth 4-manifolds

Abstract:
Some consider smooth 4-manifolds to be a mature field, which typically means its approachable yet nontrivial problems have become scarce. This is mainly due to a lack of tools. In this talk I will present a new way to depict any smooth, closed oriented 4-manifold that opens the doors to two of the most successful tools from 3-manifolds: pseudoholomorphic curves and discrete groups.

February 26, 4:30pm
Speaker: Niels Martin Moeller (Princeton)
Topic: Gluing of Geometric PDEs - Obstructions vs. Constructions for Minimal Surfaces & Mean Curvature Flow Solitons

Abstract:
For geometric nonlinear PDEs, where no easy superposition principle holds, examples of (global, geometrically/topologically interesting) solutions can be hard to come about. In certain situations, for example for 2-surfaces satisfying an equation of mean curvature type, one can generally “fuse” two or more such surfaces satisfying the PDE, as long as certain global obstructions are respected - at the cost (or benefit) of increasing the genus significantly. The key to success in such a gluing procedure is to understand the obstructions from a more local perspective, and to allow sufficiently large geometric deformations to take place. In the talk I will introduce some of the basic ideas and techniques (and pictures) in the gluing of minimal 2-surfaces in a 3-manifold. Then I will explain two recent applications, one to the study of solitons with genus in the singularity theory for mean curvature flow (rigorous construction of Ilmanen's conjectured “planosphere” self-shrinkers), and another to the non-compactness of moduli spaces of finite total curvature minimal surfaces (a problem posed by Ros & Hoffman-Meeks). Some of this work is joint w/ Steve Kleene and/or Nicos Kapouleas.

April 2, 4:30pm
Speaker: Toke Knudsen (SUNY Oneonta)
Topic: Rationales of Ancient Mathematical Methods

Abstract:
The Śulbasūtras, generally dated to 800-200 BCE, are a group of texts that provide the mathematical methods necessary for carrying out various rituals of ancient India. The texts do not seek to convince the reader that a particular formula is correct, but rather focus on providing the reader with working methods. As such, it is often not clear exactly how the authors of the texts arrived at their mathematical results. We will explore some of the mathematical statements of the texts and modern attempts at reconstructing the rationale behind them.

April 16, 4:00pm
Speaker: William Wild (SUNY Buffalo)
Topic: Eye of the Beholder: Is it Possible to Agree on What Makes an Exam Difficult?
Abstract: The level of difficulty at which students are assessed on exams can vary considerably across course sections, raising issues for both academic equity and student proficiency. This may arise when instructors vary from one another in the level of difficulty at which they intend to test. Even when instructors intend to test at the same level, however, judgments regarding what constitutes a “basic” or “difficult” question can vary widely.

This discussion will present a framework that enables instructors to rate problem difficulty in an objective manner. Used a priori, it facilitates improved control over exam design, helping instructors make purposeful and accurate choices about the difficulty profile they wish to construct. Used post hoc, it provides insight into the factors driving the exam outcomes, and by implication, student learning. The tool has validated empirically over several years use in introductory level Physics and Calculus courses.

This discussion may be of interest to:
- Instructors: seeking to gauge the level of proficiency to expect from their students, or, seeking to better understand the factors driving class performance on a given exam
- Curriculum developers: interested in achieving a greater consistency across course offerings.
- Learning Outcomes Assessors: interested in articulating outcomes that measure not only the type of student proficiency, but the level as well.

April 23, No Colloquium. Special event:
Hilton Memorial Lecture at 3pm in Science II, Room 140.
Speaker: Ralf Spatzier (University of Michigan)

April 30, 2:50pm
Dean's Lecture in Geometry and Topology.
Speaker: Karsten Grove (Notre Dame University)
Topic: Symmetry, Positive Curvature and Beyond

Abstract: Although constituting a vast extension of ancient Spherical Geometry, the beautiful class of positively curved (Riemannian) spaces is like the “Tip of the Iceberg” among all (Riemannian) spaces. Accordingly, non-symmetric positively curved spaces are known only in a few sporadic dimensions, and yet only a few obstructions to their existence are known.

In this talk, we will describe the current state of affair of the subject including tools and methods, with emphasis on the impact symmetries have had on the development during the last few decades.

May 7, 4:30pm
Speaker: Victoria Sadovskaya (Penn State)
Topic: Linear cocycles over hyperbolic systems and their periodic data

Abstract:
We consider a hyperbolic diffeomorphism f of a manifold M. A linear cocycle over f is an automorphism of a vector bundle over M that projects to f. An important example comes from the differential of f or its restriction to an invariant sub-bundle of the tangent bundle. For a trivial bundle, a linear cocycle can be viewed as a GL(d,R)-valued function on the manifold. We discuss what conclusions can be made about cocycles based on their behavior at the periodic points of f. In particular, we consider the questions when two cocycles are cohomologous and when a cocycle is conformal or isometric.

Fall 2014

October 9
Speaker: Su Yang (Chinese Academy of Sciences)
Topic: On the classification of certain 5-manifolds with fundamental group Z

Abstract:
In this talk I will give the classification of 5-manifolds with fundamental group Z and whose second homotopy group is finitely generated abelian group. As an application we obtain a criterion for 5-manifolds with fundamental group Z being a fiber bundle over the circle. The classification is also applied to classify certain knotted 3-spheres in the 5-sphere. This is a joint work with M. Kreck.

December 1
Time: 1:10 pm
Location: Science 3, Room 214
Speaker: Ruriko Yoshida (University of Kentucky)
Topic: KDETrees: Nonparametric Estimation of Phylogenetic Tree Distributions

Abstract:
While the majority of gene histories found in a clade of organisms are expected to be generated by a common process (e.g. the coalescent process), it is well-known that numerous other coexisting processes (e.g. horizontal gene transfers, gene duplication and subsequent neofunctionalization) will cause some genes to exhibit a history quite distinct from those of the majority of genes. Such “outlying” gene trees are considered to be biologically interesting and identifying these genes has become an important problem in phylogenetics.

In this talk we propose a nonparametric method of estimating distributions of phylogenetic trees, with the goal of identifying trees which are significantly different from the rest of the trees in the sample. Our method compares favorably with a similar recently-published method, featuring an improvement of one polynomial order of computational complexity (to quadratic in the number of trees analyzed), with simulation studies suggesting only a small penalty to classification accuracy. Application of our implemented software KDETrees to a set of Apicomplexa genes identified several unreliable sequence alignments which had escaped previous detection, as well as a gene independently reported as a possible case of horizontal gene transfer.

This is joint work with G. Weyenberg, P. Huggins, C. Schardl, and D. Howe.

December 1
Time: 3:30 pm
Location: WH 100E
Speaker: Angelica Cueto (Columbia University)
Topic: Non-Archimedean Combinatorics

Abstract:
Non-Archimedean analytic geometry, as developed by Berkovich, is a variation of classical complex analytic geometry for non-Archimedean fields such as p-adic numbers. Solutions to a system of polynomial equations over these fields form a totally disconnected space in their natural topology. The process of analytification adds just enough points to make them locally connected and Hausdorff. The resulting spaces are technically difficult to study but, notably, their heart is combinatorial: they can be examined through the lens of tropical and polyhedral geometry.

I will illustrate this powerful philosophy through complete examples, including elliptic curves, the tropical Grassmannian of planes of Speyer-Sturmfels, and a compactification of the well-known space of phylogenetic trees of Billera-Holmes-Vogtmann.

This talk is based on joint works with M. Haebich, H. Markwig and A. Werner.

December 2
Time: 4:30 pm
Location: WH 100E
Speaker: Ruriko Yoshida (University of Kentucky)
Topic: Markov bases for the toric homogeneous Markov Chain models

Abstract:
Discrete time Markov chains are often used in statistical models to fit the observed data from a random physical process. Sometimes, in order to simplify the model, it is convenient to consider time-homogeneous Markov chains, where the transition probabilities do not depend on the time $T$. While under the time-homogeneous Markov chain model it is assumed that the row sums of the transition probabilities are equal to one, under the toric homogeneous Markov chain (THMC) model the parameters are free and the row sums of the transition probabilities are not restricted.

In order for a statistical model to reflect the observed data, a goodness-of-fit test is applied. For instance, for the time-homogeneous Markov chain model, it is necessary to test if the assumption of time-homogeneity fits the observed data. In 1998, Diaconis-Sturmfels developed a Markov Chain Monte Carlo method (MCMC) for goodness-of-fit test by using Markov bases. A Markov basis is a set of moves between elements in the conditional sample space with the same sufficient statistics so that the transition graph for the MCMC is guaranteed to be connected for any observed value of the sufficient statistics. In algebraic terms, a Markov basis is a generating set of a toric ideal defined as the kernel of a monomial map between two polynomial rings. In algebraic statistics, the monomial map comes from the design matrix (configuration) associated with a statistical model.

In this talk we will consider a Markov basis and a Groebner basis for the toric ideal associate with the design matrix defined by the THMC model with $S \geq 2$ states without initial parameters for any time $T \geq 3$. First we will show the upper bound of the Markov degree, the degree of a minimal Markov base, of the THMC model with $S = 3$ for $T \geq 3$. In order to compute the upper bound, we use the model polytope — the convex hull of the columns of the design matrix. Here we will show the model polytope has only 24 facets for $T \geq 5$ and a complete description of the facets for $T \geq 3$. Finally, we will show a condition when the THMC with any $S \geq 2$ states for $T \geq 3$ have a square-free quadratic Groebner basis and Markov basis. One such example is the embedded discrete Markov chain (jump chain) of the Kimura three parameter model.

This is joint work with Davis Haws (IBM), Abraham Martin del Campo (IST Austria), and Akimichi Takemura (University of Tokyo).

December 3
Time: 5:00 pm (Please note the postponed time.)
Location: WH 100E
Speaker: Lucy Xia (Princeton University)
Topic: Robust Sparse Quadratic Discrimination

Abstract:
We propose a novel Rayleigh quotient based sparse quadratic dimension reduction method – named QUADRO – for analyzing high dimensional data. Unlike in the linear setting where Rayleigh quotient optimization coincides with classification, these two problems are very different under nonlinear settings. One major challenge of Rayleigh quotient optimization is that the variance of quadratic statistics involves all fourth cross-moments of predictors, which are infeasible to compute for high-dimensional applications and may accumulate too many stochastic errors. This issue is resolved by considering a family of elliptical models. Moreover, for heavy-tail distributions, robust estimates of mean vectors and covariance matrices are employed to guarantee uniform convergence in estimating nonpolynomially many parameters, even though the fourth moments are assumed. Computationally, we propose an efficient linearized augmented Lagrangian method to solve the constrained optimization problem. Theoretically, we provide explicit rates of convergence in terms of Rayleigh quotient under both Gaussian and general elliptical models. Thorough numerical results on both synthetic and real datasets are also provided to back up our theoretical results.

December 4
Time: 4:30 pm
Location: WH 100E
Speaker: Robert Haslhofer (Courant Institute, New York University)
Topic: Mean curvature flow

Abstract:
A family of hypersurfaces $M_t\subset R^{n+1}$ evolves by mean curvature flow (MCF) if the velocity at each point is given by the mean curvature vector. MCF can be viewed as a geometric heat equation, deforming surfaces towards optimal ones. If the initial surface $M_0$ is convex, then the evolving surfaces $M_t$ become rounder and rounder and converge (after rescaling) to the standard sphere $S^n$. The central task in studying MCF for more general initial surfaces is to analyze the formation of singularities. For example, if $M_0$ looks like a a dumbbell, then the neck will pinch off preventing one from continuing the flow in a smooth way. To resolve this issue, one can either try to continue the flow as a generalized weak solution or try to perform surgery (i.e. cut along necks and replace them by caps). These ideas have been implemented in the last 15 years in the deep work of White and Huisken-Sinestrari, and recently Kleiner and I found a streamlined and unified approach (arXiv: 1304.0926, 1404.2332). In this lecture, I will survey these developments for a general audience.

December 5
Time: 3:30 pm
Location: WH 100E
Speaker: Max Wakefield (US Naval Academy)
Topic: Coloring Partitions and Configuration spaces

Abstract:
There is a deep interplay between the combinatorics (matroid), algebra (cohomology or rational model), and geometry (complement) of a subspace arrangement (finite collection of subspaces in a vector space). For example if the subspaces are complex and complex codimension 1 (hyperplanes) then the Betti numbers are exactly the (unsigned) Whitney numbers of the first kind on the intersection lattice. Subspace arrangements of the braid arrangement can be enumerated by partitions. It turns out that the Whitney numbers of these subspace arrangements can be found by looking at a generalized chromatic polynomial of the associated partitions. Unfortunately, these Whitney numbers do not give the Betti numbers of the complement and finding a closed formula for these Betti numbers is not known. However, using tools from rational homotopy theory we can show that certain classes of these arrangements are rationally formal and non-formal. At the end we will construct a new differential graded algebra which presents a kind of model for the collection of all k-equal arrangements (configuration spaces where k-1 points can collide) which gives hints at a nice presentation for the cohomology and the Betti numbers.

December 5
Time: 5:00 pm (Please note the postponed time.)
Location: WH 100E
Speaker: Quefeng Li (Princeton University)
Topic: Robust Estimation of High-Dimensional Mean Regression

Abstract:
Data subject to heavy-tailed errors are commonly encountered in various scientific fields, especially in the modern era with explosion of massive data. To address this problem, procedures based on quantile regression and Least Absolute Deviation (LAD) regression have been developed in recent years. These methods essentially estimate the conditional median (or quantile) function. They can be very different from the conditional mean functions when distributions are asymmetric and heteroscedastic. How can we efficiently estimate the mean regression functions in ultra-high dimensional setting with existence of only the second moment? To solve this problem, we propose a penalized Huber loss with diverging parameter to reduce biases created by the traditional Huber loss. Such a penalized robust approximate quadratic (RA-quadratic) loss will be called RA-Lasso. In the ultra-high dimensional setting, where the dimensionality can grow exponentially with the sample size, our results reveal that the RA-lasso estimator produces a consistent estimator at the same rate as the optimal rate under the light-tail situation. We further study the computational convergence of RA-Lasso and show that the composite gradient descent algorithm indeed produces a solution that admits the same optimal rate after sufficient iterations. As a byproduct, we also establish the concentration inequality for estimating population mean when there exists only the second moment. We compare RA-Lasso with other regularized robust estimators based on quantile regression and LAD regression. Extensive simulation studies demonstrate the satisfactory finite-sample performance of RA-Lasso.

December 8
Time: 5:00 pm
Location: WH 100E
Speaker: Daniel Sewell (University of Illinois at Urbana-Champaign)
Topic: Latent space models for dynamic networks

Abstract:
Dynamic networks are used in a variety of fields to represent the structure and evolution of the relationships between entities. We present a model which embeds longitudinal network data as trajectories in a latent Euclidean space. A Markov chain Monte Carlo algorithm is proposed to estimate the model parameters and latent positions of the actors in the network. The model yields meaningful visualization of dynamic networks, giving the researcher insight into the evolution and the structure, both local and global, of the network. The model handles directed or undirected edges, easily handles missing edges, and lends itself well to predicting future edges. Further, a novel approach is given to detect and visualize an attracting influence between actors using only the edge information. We use the case-control likelihood approximation to speed up the estimation algorithm, modifying it slightly to account for missing data. We apply the latent space model to data collected from a Dutch classroom, and cosponsorship network collected on members of the U.S. House of Representatives, illustrating the usefulness of the model by making insights into the networks.

December 10
Time: 5:00 pm
Location: WH 100E
Speaker: Zuofeng Shang (Purdue University)
Topic: Nonparametric Bernstein-von Mises Phenomenon: A Tuning Prior Perspective

Abstract:
Statistical inference on infinite-dimensional parameters in Bayesian framework is investigated. The main contribution of our work is to demonstrate that nonparametric Bernstein-von Mises theorem can be established in a very general class of nonparametric regression models under the novel tuning priors. Surprisingly, this type of prior connects two important classes of statistical methods: nonparametric Bayes and smoothing spline at a fundamental level. The association with smoothing spline facilitates both theoretical analysis and applications for nonparametric Bayesian inference. For example, the selection of a proper tuning prior can be easily done through generalized cross validation, which can be well implemented by existing R packages.

This is a joint work with Guang Cheng (Purdue).

December 11
Time: 2:50 pm
Location: WH 100E
Speaker: Mark Hagen (University of Michigan)
Topic: Coarse, flexible, and rigid structures in geometric group theory

Abstract:
Geometric group theory is partly the study of groups arising naturally in geometry and topology: this includes fundamental groups of interesting spaces (e.g. 3-manifold groups) or groups of symmetries (isometries, homeomorphisms, etc.) of interesting spaces (e.g. mapping class groups). Geometric group theory is also the study of groups as geometric objects in their own right.

This talk deals with three viewpoints from which a group can be analyzed, and the interplay between these. First, the source of many questions in geometric group theory is the topological viewpoint, in which spaces are distinguished up to homeomorphism, homotopy equivalence, etc. Once one has isolated the fundamental group of one's space, and found generators, the natural geometry becomes “coarse”, and things are generally true up to a relation called “quasi-isometry”. Often, it is desirable to realize the group as a group of automorphisms of some very specific, rigid combinatorial structure; this is the third viewpoint. I will discuss examples of how each of these approaches can naturally lead to and interact with the others.

I will conclude with a brief discussion of very recent joint work with J. Behrstock and A. Sisto, in which we define a class of spaces that includes mapping class groups, many cubical groups (which will have been defined), and most 3-manifold groups, and build tools to study the coarse geometry of such spaces from a common perspective.

December 12
Time: 5:00 pm
Location: WH 100E
Speaker: Michael Dobbins (Center for Geometry and its Applications, Pohang University of Science and Technology)
Topic: A Point in a $nd$-Polytope is the Barycenter of $n$ Points in its $d$-Faces.

Abstract:
In this talk I show that it is always possible to find $n$ points in the $d$-dimensional faces of a $nd$-dimensional convex polytope $P$ so that their center of mass is a target point in $P$. Equivalently, the $n$-fold Minkowski sum of the polytope's $d$-skeleton is the polytope scaled by $n$. This verifies a conjecture by Takeshi Tokuyama.

December 15
Time: 1:45 pm
Location: WH 100E
Speaker: Michael Jablonski (University of Oklahoma)
Topic: Non-compact, homogeneous Einstein spaces.

Abstract: For over a century, Einstein metrics have remained of core interest in modern geometry. On a homogeneous space the Einstein condition reduces to a collection of polynomials and so, in principal, such spaces should be easy to understand and classify. However, the reality is much more complicated and no classification exists in either the compact or non-compact settings. In this talk, we present the current state of knowledge on the classification of non-compact, homogeneous Einstein spaces.

Colloquium

2026-05-04T09:09:42-04:00

Colloquium

Unless stated otherwise, colloquia are scheduled for Thursdays 4:00-5:00pm in WH-100E with refreshments served from 3:45-4:00 pm in WH-102.

Organizers: Michael Dobbins, Vladislav Kargin, Cary Malkiewich, Adrian Vasiu, and Emmett Wyman

Fall 2025

Thursday Nov 6 4:00-5:00pm, WH-100E
Speaker: Andrew Obus (CUNY)
Topic: The lifting problem for covers of curves, particularly its group-theoretical aspects

Abstract: Whenever a mathematical object is given in characteristic p, one can ask whether it is the reduction, in some sense, of an analogous structure in characteristic zero. If so, the structure in characteristic zero is called a “lift” of the structure in characteristic p. The most famous example is Hensel's Lemma about lifting solutions of polynomials in Z/p to solutions in the p-adic integers Z_p.

The “lifting problem” we consider is more geometric: given a smooth curve X in characteristic p with an action of a finite group G, is there a curve in characteristic zero with G-action that reduces to X? Unsurprisingly, the answer is related to the group theory of G (for instance, if p does not divide |G| or if G is cyclic, then the curve with the G-action always lifts, but if G has an abelian, non-cyclic, non-p-subgroup that fixes a point on X, then the curve does not lift with the action). After giving an introduction to the lifting problem and some examples, we will discuss well-established ways that the problem interacts with group theory, as well as more recent advances relating the problem to representation theory.

Spring 2026

March 13th
PETER HILTON MEMORIAL LECTURE
SPECIAL TIME AND LOCATION: March 13, 3:30pm, Alumni Lounge at Old O'Connor Hall
Speaker: Martin Bridson (University of Oxford)
Title: Chasing finite shadows of infinite groups through geometry

Abstract: There are many situations in geometry or elsewhere in mathematics where it is natural or convenient to explore infinite groups of symmetries via their actions on finite objects. But how hard is it find these finite manifestations and to what extent does the collection of all such actions determine the infinite group?

In this colloquium, I will sketch some of the rich history of such problems and then describe some of the great advances in recent years. I'll describe pairs of distinct groups that have the same finite images and I'll sketch the proof of some “profinite rigidity results”, i.e. theorems showing that in certain circumstances one can identify an infinite group if one knows its set of finite images.

Thursday May 7, 2026 2:45-3:45pm, WH-100E
Speaker: Isabella Novik (University of Washington)
Topic: Lower bounds on face numbers

Abstract: In this talk, I will discuss several approaches to studying the face numbers of simplicial complexes, with a particular focus on obtaining lower bounds. These approaches include classical rigidity theory, Stanley-Reisner rings, and higher‑dimensional stress spaces. I will also describe a number of both classical and recent results on face numbers that have been obtained using these tools.

Archive:

seminars:comb

2020-01-29T15:14:10-04:00

This page has been moved, the new location is Combinatorics Seminar.

Data Science Seminar

2026-04-24T11:30:48-04:00

Data Science Seminar

The Data Science Seminar is evolved from the former Statistical Machine Learning Seminar which covered topics in statistical theory that was important for machine learning research as well as development and applications of machine learning techniques in interdisciplinary research. The scope of the Data Science Seminar has been broadened to facilitate dialogue among different communities in the data science circle.

It is listed as course MATH 568.

Location: Whitney 100E (See the directions to the department)
Time: Tuesday 12:15 pm –1:15 pm
Organizers: Minghao Rostami and Minjie Wang

Spring 2026

February 10, 2026
Speaker: Dr. Yizeng Li (Department of Biomedical Engineering at Binghamton University)
Topic: Multiphase Continuum Models for Cell Migration.
Abstract

April 14, 2026
Speaker: Dr. Yang Feng (New York University)
Topic: Transfer and Multi-task Learning: Statistical Insights for Modern Data Challenges.
Abstract

April 21, 2026
Speaker: Dr. Yiqun T. Chen (Johns Hopkins University)
Topic: AI for (Bio)statistics and (Bio)statistics for AI.
Abstract

April 28, 2026
Speaker: Joyce Zhou (Cornell University)
Topic: Using Natural Language to Steer Recommendation.
Abstract

May 5, 2026
Speaker: Dr. Jun Yan (University of Connecticut)
Topic: Joint Observation-Constrained Climate Projections for Spatial Fields Using Hierarchical Emergent Constraints.
Abstract

Fall 2025

September 9, 2025
Speaker: Dr. Nancy Guo (School of Computing at Binghamton University)
Topic: AI-empowered precision medicine.
Abstract

October 21, 2025
Speaker: Dr. Fei Xue (Purdue University)
Topic: Statistical Methods for Mobile Health Data.
Abstract

October 28, 2025
Speaker: Dr. Yufeng Liu (University of Michigan)
Topic: Low-Rank Online Dynamic Assortment with Dual Contextual Information.
Abstract

November 18, 2025
Speaker: Dr. Bingxin Zhao (University of Pennsylvania)
Topic: Resampling-based pseudo-training in genomic predictions.
Abstract

Spring 2025

February 25, 2025
Speaker: Dr. Zhaohan Xi (School of Computing at Binghamton University)
Topic: From Text to Impact: Large Language Models as Responsible Cross-Disciplinary Copilots.
Abstract

March 18, 2025
Speaker: Dr. Kun Chen (University of Connecticut)
Topic: Hybrid and Integrative Learning for Rare Event Modeling with EHR Data.
Abstract

April 8, 2025
Speaker: Dr. Lingzhou Xue (Pennsylvania State University)
Topic: Federated On-Policy Reinforcement Learning.
Abstract

April 15, 2025
Speaker: Dr. Min Xu (Rutgers University - New Brunswick)
Topic: Optimal Convex M-Estimation via Score Matching.
Abstract

Fall 2024

October 15, 2024
Speaker: Dr. Yang Ning (Cornell University)
Topic: Estimation and Inference in Multivariate Response Regression with Hidden Variables.
Abstract

October 22, 2024
Speaker: Dr. JoonHwan Cho (Department of Economics at Binghamton University)
Topic: Testing for exogenous participation in ascending auction with unobserved heterogeneity.
Abstract

November 5, 2024
CANCELLED AND POSTPONED
Speaker: Dr. Yunlong Feng (SUNY Albany)
Topic: Understanding robust loss functions in machine learning.
Abstract

November 12, 2024
Speaker: Ben Jones
Topic: An Integrated Experimental and Modeling Approach to Design Rotating Algae Biofilm Reactors (RABRs) via Optimizing Algae Biofilm Productivity, Nutrient Recovery, and Energy Efficiency.
Abstract

December 3, 2024
Speaker: Dr. Yingxue Zhang (School of Computing at Binghamton University)
Topic: Leveraging Unlabeled Data in Offline Reinforcement Learning.
Abstract

Spring 2024

January 23, 2024
Speaker: Yili Zhang (MathWorks)
Topic: Low-Code Machine Learning in SIMULINK & MATLAB APPS.
Abstract

March 26, 2024
Speaker: Dr. Zeyu Ding (Department of Computer Science at Binghamton University)
Topic: Differential Privacy in Practice: How the US Government Protects Your Sensitive Information in the 2020 Census.
Abstract

Fall 2023

September 26, 2023
Speaker: Dr. Marianthi Markatou (SUNY University at Buffalo)
Topic: Distances and their role in statistical inference.
Abstract

October 3, 2023
Speaker: Dr. HaiYing Wang (University of Connecticut)
Topic: Rare Events Data and Maximum Sampled Conditional Likelihood.
Abstract

October 10, 2023
Speaker: Dr. Yiming Ying (SUNY University at Albany)
Topic: Interplay between Generalization and Optimization via Algorithmic Stability.
Abstract

October 17, 2023
Speaker: Dr. Peter D. Hoff (Duke University)
Topic: Core Shrinkage Covariance Estimation for Matrix-variate Data.
Abstract

October 24, 2023
Speaker: Dr. Luis Carvalho (Boston University)
Topic: Deviance Matrix Factorization.
Abstract

October 31, 2023
Speaker: Dr. Ruiqi Liu (Texas Tech University)
Topic: Estimation and Hypothesis Testing of Derivatives in Smoothing Spline ANOVA Models.
Abstract

November 14, 2023
Speaker: Dr. Jiguo Cao (Simon Fraser University)
Topic: Machine Learning for Functional Data.
Abstract

November 28, 2023
Speaker: Dr. Li Zhang (University of California San Francisco)
Topic: NAIR Software: Unlocking the Immune System's Secrets by Network Analysis and Advanced Machine Learning.
Abstract

December 5, 2023
Speaker: Dr. Xuexia Wang (Florida International University)
Topic: Genetic Association Test and Risk Prediction Modeling for Cardiomyopathy in Cancer Survivors.
Abstract

Spring 2023

February 14, 2023
Speaker: Dr. Yuan Fang (Binghamton University)
Topic: Clustering disease trajectories: statistical method applications and evaluation.
Abstract

February 21, 2023
Speaker: Dr. Xuming He (University of Michigan)
Topic: How Good is Your Best Selected Subgroup.
Abstract

February 28, 2023
Speaker: Dr. James D. Wilson (University of San Francisco)
Topic: The Political Brain: Associations of Tasked-based Functional Connectivity Networks and Political Ideology.
Abstract

March 7, 2023
Speaker: Dr. Sijian Wang (Rutgers, The State University of New Jersey)
Topic: Dynamic Attention-Based Functional Data Analysis.
Abstract

March 14, 2023
Speaker: Dr. Holger Dette (Ruhr-Universitaet Bochum)
Topic: Functional data analysis on Banach spaces.
Abstract

March 21, 2023
Speaker: Dr. Hamsa Bastani (University of Pennsylvania)
Topic: Efficient and targeted COVID-19 border testing via reinforcement learning.
Abstract

March 28, 2023
Speaker: Dr. Jie Peng (UC Davis)
Topic: Statistical methods for diffusion MRI.
Abstract

April 11, 2023
Speaker: Dr. Chris Haines (Internal)
Topic: Independent Spacings Theorem with a Maximum Product Spacings Estimation Application.
Abstract

April 18, 2023
Speaker: Dr. Konstantin G. Arbeev (Duke University)
Topic: How Good is Your Best Selected SubgroupStochastic process models: Bringing biology to statistics to advance research on aging.
Abstract

April 27, 2023
Speaker: Dr. Runze Li (Penn State University)
Topic: Model-Free Conditional Feature Screening with FDR Control.
Abstract

Fall 2022

September 20, 2022
Speaker: Dr. Soumik Banerjee (Internal)
Topic: Likelihood-based Approach for Testing the Homogeneity of Risk Difference in a Multicenter Randomized Clinical Trial.
Abstract

October 4, 2022
Speaker: Dr. Chao Huang (Florida State University)
Topic: Shape-on-Scalar Regression Models: Going Beyond Prealigned Non-Euclidean Responses.
Abstract

October 25, 2022
Speaker: Dr. John Stufken (George Mason University)
Topic: Musings on Subdata Selection.
Abstract

November 1, 2022
Speaker: Dr. Jinchi Lv (University of Southern California)
Topic: High-Dimensional Knockoffs Inference for Time Series Data.
Abstract

November 8, 2022
Speaker: Dr. Rui Song (North Carolina State University)
Topic: On statistical inference for sequential decision making.
Abstract

November 15, 2022
Speaker: Dr. Rachael Hageman Blair (University at Buffalo)
Topic: Harnessing stability estimation for module detection, clustering, and ensemble clustering.
Abstract

December 6, 2022
Speaker: Dr. Alexander Franks (University of California, Santa Barbara)
Topic: Sensitivity to Unobserved Confounding in Studies with Factor-structured Outcomes.
Abstract

Spring 2022

Apr. 5, 2022
Speaker: Dr. Soumik Banerjee (Internal)
Topic: Multistage Minimum Risk Point Estimation (MRPE) with First-Order and Second-Order Asymptotic Properties.
Abstract

Apr. 26, 2022
Speaker: Dr. Krishnakumar Balasubramanian (The University of California, Davis)
Topic: Fractal Gaussian Networks: A sparse random graph model based on Gaussian Multiplicative Chaos.
Abstract

May 3, 2022
Speaker: Dr. Hongtu Zhu (University of North Carolina)
Topic: Challenges in Biobank-scale: Imaging Genetics and Beyond.
Abstract

May 10, 2022
Speaker: Dr. Yao Zheng (University of Connecticut)
Topic: Tensor methods for high-dimensional time series modeling.
Abstract

Fall 2021

Sep. 14, 2021
Speaker: Dr. Matthew Reimherr (Pennsylvania State University)
Topic: KNG - A New Mechanism for Data Privacy.
Abstract

Sep. 21, 2021
Speaker: Dr. Haoda Fu (Eli Lilly and Company)
Topic: Our Recent Development on Cost Constraint Machine Learning Models.
Abstract

Sep. 28, 2021
Speaker: Dr. Eric F. Lock (University of Minnesota)
Topic: Bidimensional Linked Matrix Decomposition for Pan-Omics Pan-Cancer Analysis.
Abstract

Oct. 19, 2021
Speaker: Dr. Damla Senturk (UCLA)
Topic: Multilevel Modeling of Spatially Nested Functional Data: Spatiotemporal Patterns of Hospitalization Rates in the U.S. Dialysis Population.
Abstract

Oct. 26, 2021
Speaker: Dr. Giles Hooker (UC Berkeley)
Topic: There is No Free Variable Importance: Traps in Interpreting Black Box Functions.
Abstract

Nov. 2, 2021
Speaker: Dr. Yuanjia Wang (Columbia University)
Topic: Machine Learning Approaches for Optimizing Treatment Strategies for Mental Disorders.
Abstract

Nov. 9, 2021
Speaker: Dr. Megan Johnson (Internal)
Topic: The Interconnectivity Vector and the Betti Sequence: Finite-Dimensional Vector Representations of Persistent Homology.
Abstract

Nov. 16, 2021
Speaker: Dr. Cen Wu (Kansas State University)
Topic: Robust Bayesian variable selection for gene-environment interactions.
Abstract

Nov. 30, 2021
Speaker: Dr. Antonio Linero (University of Texas at Austin)
Topic: Bayesian Decision Tree Ensembling Strategies for Nonparametric Problems.
Abstract

Dec. 7, 2021
Speaker: Dr. Annie Qu (University of California Irvine)
Topic: Correlation Tensor Decomposition and Its Application in Spatial Imaging Data.
Abstract

Spring 2021

Mar. 02, 2021
Speaker: Brian Franczak (MacEwan University)
Topic: On using mixtures of shifted asymmetric Laplace distributions for model-based classification.
Abstract

Mar. 23, 2021
Speaker: Adam Ciarleglio (The George Washington University)
Topic: Multiple imputation in functional regression with applications to EEG data in a depression study.
Abstract

March 30, 2021
Speaker: Wenshu Dai (Binghamton University)
Topic: Finite Mixtures of Regression Models and Finite Mixtures of Regression Models with Concomitant Variables for Clustering Microbiome Data.
Abstract

April. 13, 2021
Speaker: Nalini Ravishanker (University of Connecticut)
Topic: Biclustering Approaches for High-Frequency Time Series.
Abstract

April. 27, 2021
Speaker: Melody Ghahramani (The University of Winnipeg)
Topic:Time Series Regression for Zero-Inflated and Overdispersed Count Data: A Functional Response Model Approach.
Abstract

May 04, 2021
Speaker: Zhou Wang (Binghamton University)
Topic: Multiclass Anomaly Detector: the CS++Support Vector Machine
Abstract

Fall 2020

Oct. 20, 2020
Speaker: Sumanta Basu (Cornell University)
Topic: Measuring Systemic Risk with Graphical Models of Time Series Data.
Abstract

Oct. 27, 2020
Speaker: Yuan Luo (Northwestern University)
Topic: A Multidimensional Precision Medicine Approach Identifies an Autism Subtype Characterized by Dyslipidemia
Abstract

Nov. 3, 2020
Speaker: Shaofei Zhao (Binghamton University)
Topic: Distribution-free and nonparametric multivariate feature screening via measure transportation for high dimensional response and predictor variables.
Abstract

Spring 2020

Mar. 24, 2020
CANCELLED AND POSTPONED
Speaker: Sumanta Basu (Cornell University)

April. 21, 2020
Speaker: Liang Li, Yunhui Liu and Han Zhang
Topic: Capstone Project: Factors Affecting PhD Student Success.
Abstract

April. 21, 2020
CANCELLED AND POSTPONED
Speaker: Adam Ciarleglio (George Washington University)

May. 05, 2020
CANCELLED AND POSTPONED
Speaker: Yuan Luo (Northwestern University)

—

Fall 2019

See the schedule of the Interdisciplinary Dean's Speaker Series in Data Science.

Oct. 9, 2019 (special day and time)
Interdisciplinary Dean's Speaker Series in Data Science
Speaker: Joseph W Hogan (Brown University)
Topic: Using Electronic Health Records Data for Predictive and Causal Inference About the HIV Care Cascade
Abstract

Oct. 11, 2019 (Special Date and time)
Speaker: Paul McNicholas (McMaster University)
Topic: Clustering Higher-Order Data
Abstract

Oct. 22, 2019
Speaker: Gen Li (Columbia University)
Topic: Integrative multi-view regression: Bridging group-sparse and low-rank models.
Abstract

Oct. 24, 2019 (1:15 pm, stat seminar time)
Speaker: Doug Turnbull (Ithaca College)
Topic: TBA
Abstract

Oct. 29, 2019
Speaker: Wangshu Tu (Binghamton University)
Topic: A family of mixture models for biclustering
Abstract

Nov. 5, 2019
Speaker: Wangshu Tu (Binghamton University)
Topic: Non existence of fixed sample estimator for prescribed proportional closeness
Abstract

Nov. 8, 2019
Interdisciplinary Dean's Speaker Series in Data Science
Speaker: Andrew Gordon Wilson (New York University; Courant Institute of Mathematical Sciences)
Topic: How do we build models that learn and generalize?
Abstract

Nov. 12, 2019
Speaker: Kexuan Li (Binghamton University)
Topic: A Hausman test for the presence of market microstructure noise in high frequency data
Abstract

Nov. 19, 2019
Interdisciplinary Dean's Speaker Series in Data Science
Time: 10am-11:30am
Location: UUW325
Speaker: Arthur Spirling (New York University; Politics and Data Science)
Topic: Word Embeddings: What works, what doesn’t, and how to tell the difference for applied research
Abstract

Nov. 26, 2019
Speaker: Wei Yang (Binghamton University)
Topic: Random Covariance Matrix and the Marchenko-Pastur law
Abstract

—

Spring 2019

March 05, 2019
Speaker: Rebecca Nugent (Carnegie Mellon University)
Topic: Before Teaching Data Science, Let’s First Understand How People Do It
Abstract

March 12, 2019
Speaker: Subhadeep (Deep) Mukhopadhyay (Temple University)
Topic: Graph Data Science
Abstract

March 26, 2019
The Dean's Speaker Series in Statistics and Data Science
Speaker: Regina Y. Liu (Rutgers University)
Topic: Fusion Learning: Efficient Combination of Inferences from Diverse Data Sources
Abstract

April 9, 2019
Speaker: David Hunter (Pennsylvania State University)
Topic: Multivariate Nonparametric Mixture Models
Abstract

April 16, 2019
The Dean's Speaker Series in Statistics and Data Science
Speaker: David Madigan (Columbia University)
Topic: Towards honest inference from real-world healthcare data
Abstract

April 23, 2019
Speaker: Daphney-Stavroula Zois (SUNY Albany)
Topic: Spatiotemporal Quickest Change Detection for Traffic Accident Nowcasting
Abstract

April 30, 2019
Speaker: Lin Yao (Binghamton University)
Topic: Dissertation Defense - JAMES-STEIN-TYPE OPTIMAL WEIGHT CHOICE FOR FREQUENTIST MODEL AVERAGE ESTIMATOR
Special time and location: 3:30 pm at OR 100D
Abstract

—

Fall 2018

October 9, 2018
The Dean's Speaker Series in Statistics and Data Science
Speaker: Sidney Resnick (Cornell University)
Location: Old Champlain Atrium (unusual location)
Topic: Fitting the Linear Preferential Attachment Model for Social Network Growth
Abstract

October 23, 2018
The Dean's Speaker Series in Statistics and Data Science
Speaker: David Ruppert (Cornell University)
Topic: Density Estimation with Noisy Data
Abstract

November 13, 2018
Speaker: Yudong Chen (Cornell University)
Topic: Byzantine-Robust Distributed Learning with Non-converxity
Abstract

—-

Spring 2018

January 30
Speaker: Qiqing Yu (Binghamton University)
Topic: Identifiability Conditions For The Linear Regression Model Under Right Censoring
Abstract

February 20
Speaker: Yinsong Chen (Binghamton University)
Topic:The Conductance and Mixing Time
Abstract

March 13
Speaker: Jiexin Duan (Purdue University)
Topic: Large-Scale Nearest Neighbor Classification with Statistical Guarantee
Abstract

March 20
Speaker: Yuan Fang (Binghamton University)
Topic: Bayesian Approach to Parameter Estimation for the mixtures of Multivariate Normal Inverse Gaussian Distributions
Abstract

April 10
Speaker: Leila Setayeshgar (Providence College)
Topic: Large Deviations for a Class of Stochastic Semilinear Partial Differential Equations
Abstract

April 17
Speaker: Wenbo Wang (Binghamton University)
Topic: A look at distance-weighted discrimination
Abstract

April 24
Speaker: Haomiao Meng (Binghamton University)
Topic: Multicategory Angle-based Large-margin Classification
Abstract

May 1
Speaker: Chen Liang (Binghamton University)
Topic: Goodness of fit tests for clustered spatial point processes
Abstract

—-

Fall 2017

September 19
Speaker: Dan Yang (Rutgers University)
Topic: Autoregressive Model for Matrix Valued Time Series
Abstract

September 26
Speaker: Ji Meng Loh (New Jersey Institute of Technology)
Topic: Single-index model for inhomogeneous spatial point processes
Abstract

November 14
Speaker: Weijie Su (Wharton School of the University of Pennsylvania)
Topic: TBA
Abstract

Capstone Seminar

2017-10-27T14:42:48-04:00

Capstone Seminar

Fall 2017

September 13
Time: 1:10–1:25 pm
Speaker: Wangshu Tu
Title: Parameters Selection and Comparison in Guassian Kernel SVM

Abstract: In SVM(Support Vector Machines), it is not so clear that which kernel to choose and how to select proper parameter in kernel function. One of them: Gaussian radial basis function(RBF) is very popular because of only single parameter needs to be determined. In this short talk, it will present different results of applying RBF kernel in binary classification case–Gender Recognition by Voice, with different pairs of (C, r), where C is a regularization parameter to constrain the range of Lagrangian coefficients in dual function F_D, r is reciprocal of single parameter sigma^2 in RBF kernel. For each pair (C, r), compute 10-folder Cross Validation(CV/10) misclassification rate, and it indicates the rate will decrease when C or r increases. The smallest CV/10 misclassification rate among all pairs of (C, r) is also better than LDA and classification tree.

September 13
Time: 1:25–1:40 pm
Speaker: Xiang Wang
Title: Use LDA and QDA To Discriminate Diabetes Data

Abstract: Diabetes data resulted from a study conducted at the Stanford Clinical Research Center of the relationship between the three clinical classifications and five measurements for 145 instances. It helps the diagnosis and appropriate treatment to the diabetes patients.

We draw a scatterplot matrix of all five variables representing the problematic multivari- ate Gaussian distributions, and the assumption of equal covariance matrices is inappropriate, which play the negative roles in LDA(Linear Discriminant Analysis) and QDA(Quadratic Discriminant Analysis). We use LDA and QDA to discriminate the clinical classifications by the five variables, then draw the 2D-scatter-plot of the first two discriminating functions to show that LDA is subject to outliers, but QDA relatively improves the classification by nonlinear discrimination.

The misclassification rates in the leave-one-out cross-validation are 11% for LDA and 9.7% for QDA. The fitting for 145 instances indicates LDA and QDA can be quite flexible.

September 13
Time: 1:40–1:55 pm
Speaker: Joshua Rovou
Title: Managing Multinomial Data: Using Aids and Examples from Dr. Ganggang Xu and Julian Faraway's “Generalized Linear Models”

Abstract: Multinomial data requires careful thought to properly analyze. There are various forms that multinomial data can take that can easily be misclassified, leading to false conclusions. Understanding when, why, and how to recognize and apply these forms and their assumptions allows a data scientist to be mindful of best practices when fitting multinomial models. This paper provides an overview, suggestions, and examples of classifying multinomial data. In addition, this paper provides discussions on applying the appropriate assumptions and models in the R programming language, using the problem and data sets in Faraway’s “Generalized Linear Models” Chapter 5 as case studies. These examples seek to illustrate best practices when dealing with multinomial data for the student data scientist.

September 13
Time: 1:55–2:10 pm
Speaker: Hao Wang
Title: Forest Cover Type prediction

Abstract: This report is based on the data from UCI machine learning website and the raw data was determined from the US Forest Service (USFS) Region 2 Resource Information System (RIS). There are totally 581,012 observations in this data set and each observation is a 30m x 30m cell containing 54 attributes, which consist of 10 quantitative variables, 4 binary wilderness areas and 40 binary soil type variables. The forest cover type is basically a classification problem. By looking from the previous work from Blackard's investigation on this topic (1), we think that we could try using several classifiers such as SVM,K-NN, decision tree and gradient boost model in modeling to increase the accuracy in prediction.

September 20
Time: 1:10–1:25 pm
Speaker: Xiaolin Tang
Title: MODEL SELECTION

Abstract: TBA.For a dataset, we can come up with plenty of models, however, there should be a best one through all of them. Therefore, we need some criteria to evaluate the model, and a method to find out it, here comes model selection.

My presentation will be presented with two parts: first, I will give a quickly review of some basic concepts, including the criteria of a good model and the method to do model selection. Second, I will display several questions of chapter 10 in “Faraway I”(Linear Models with R, second Edition) and talk about the solutions. For Question 10.5, it's about how outliers influence the result of model selection, we use the data with and without the outliers to do model selection separately, and our conclusion is we can obtain the same model, however, the estimated coefficient will be very different for these two models.

September 20
Time: 1:25–1:40 pm
Speaker: Shaofei Zhao
Title: Use Dataset aatemp to Predict Temperature in the Future

Abstract: Dataset aatemp is from the U.S. Historical Climatology Network. They are the annual mean temperatures in Ann Arbor, Michigan going back about 150 years. The data contains 115 observations on two variables, year is the year from 1854 to 2000, temp is the annual mean temperature. Our motivation is that by analyzing the data, we may give a reasonable prediction of the mean temperature in 2020. To analyze, we need firstly check the assumptions of error terms, such as constant variance, normality and outliers. Then we proceed linear regression considering response transformation and predictor transformation, in the long talk we may also consider some methods of time series to achieve a better model. At the moment, we use a model with 3-degree polynomials, which can increase our R^2 from 0.05 to 0.13, and base on this model, we may predict the temperature of 2020 is about 46 degrees.

September 20
Time: 1:40–1:55 pm
Speaker: McInroy,Alexander
Title: Binomial Regression and its Applications in Medical Diagnosis

Abstract: Regression analysis remains one of the most directly applicable tools in the field of statistics and machine learning today. In particular, binomial regression is especially useful for diagnosing medical conditions where a test leads to a positive or negative result. This talk will use examples from Extending the Linear Model with R, Faraway to illustrate this. Technical topics covered will be data analysis, data interpretation, model selection, and model prediction. Furthermore, selecting appropriate cutoff points for a binomial model’s response will be discussed, since mitigating type I and type II error is a subjective balancing act depending on the situation..

September 20
Time: 1:55–2:10 pm
Speaker: Schepis,Michael
Title: Classification Methodology

Abstract: One of the primary motivations of analyzing data is our ability to accurately categorize it. This is the purpose of classification; Different statistical techniques allow us to divide categorical data into relevant groups, however this ability is only as useful as our ability to interpret these groups in a meaningful way. Most data of interest cannot be perfectly separated without error, so it is also crucial to analyze mis-classification rates with each of these statistical tools. In this talk, we will provide relevant examples from Alan J. Izenman's Multivariate Statistical Techniques and discuss the tools of Linear Discriminant analysis, Bayes Rule, and Quadratic Discriminant Analysis and when each of those techniques would be most appropriate and how to create a confusion matrix to examine the accuracy of our findings..

September 27
Time: 1:10–1:25 pm
Speaker: Yanwei Jiang
Title: Lineal Model diagnostics in error terms

Abstract: When a linear model has been constructed, we need to check whether the model is appropriate and do some adjustment if necessary, this process is called model diagnostics. There are several aspects that we need to consider, in this short talk, we will focus on the error term of linear model. The model assumptions of the error term are constant variance, normality and independent. We will introduce some common and useful tools such as residual plot, qq-plot and other tools, both in theory and method. Some examples will also be demonstrated to show how these tools work.

September 27
Time: 1:25–1:40 pm
Speaker: Yifei Zeng
Title: Application of Multidimensional Scaling

Abstract: Multidimensional Scaling (mds) is a method to reduce high-dimensional-data to low-dimensional-data, most likely to 2-dimensions or 3-dimensions which is the dimension that human can visualize. Also mds can also be used to cluster different groups of data. In this talk, we will discuss different types of mds and apply them to a data set from Izenman. The data set contains the distance between 48 cities in UK which forms a dissimilarity matrix. We will compare different types of mds with respect to visualization, computation, and fits of data etc. Also within a distinct type of mds, we will talk about how to choose the dimension so when we are reducing the dimension, we will still get enough information from data. We will finally present the map of the 48 cities and compare with the mds’ map to see the fit of mds.

September 27
Time: 1:40–1:55 pm
Speaker: Gang Cheng
Title: Problems with the error term

Abstract: In ordinary least square regression, we assume the error term $\epsilon$ is independent and identically distributed. Furthermore, in order to carry out the usual statistical inference, we also assume the error term are normally distributed. However, in many cases, this assumption always violated and we have to consider alternatives. (i) When the errors are dependent, like time series, we use \emph{Generalized Least Squares}(GLS); (ii) When the errors are independent, but not identically distributed, we can use \emph{Weighted Least Squares}. (iii)When the errors are not normally distributed, we can use \emph{Robust Regression}. In this talk, I will mainly focus on the theory part of these regressions and one or two example(s) of these.

September 27
Time: 1:55–2:10 pm
Speaker: Chenxi Wang
Title: Linear Model with categorical predictors

Abstract: My topic for seminar is focused on Chapter 14 of the book Linear Models of R by Faraway. This chapter mainly talks about categorical predictors for linear regression models. I will give brief talk about basic concepts related to categorical predictors. Also, together with exercises at the end of this chapter, I will give specific examples on how to deal with categorical factors in regression analysis in practical world.

October 4
Time: 1:10–1:40 pm
Speaker: Hao Wang
Title: Forest Cover Type prediction

Abstract: This report is based on the data from UCI machine learning website and the raw data was determined from the US Forest Service (USFS) Region 2 Resource Information System (RIS). There are totally 581,012 observations in this data set and each observation is a 30m x 30m cell containing 54 attributes, which consist of 10 quantitative variables, 4 binary wilderness areas and 40 binary soil type variables. The forest cover type is basically a classification problem. By looking from the previous work from Blackard's investigation on this topic (1), we think that we could try using several classifiers such as SVM,K-NN, decision tree and gradient boost model in modeling to increase the accuracy in prediction.

October 4
Time: 1:40–2:10 pm
Speaker: Chenxi Wang
Title: Linear Models with Categorical Predictors

Abstract: My topic for seminar is focused on Chapter 14 of the book Linear Models of R by Faraway. This chapter mainly talks about categorical predictors for linear regression models. I will give brief talk about basic concepts related to categorical predictors. Also, together with exercises at the end of this chapter, I will give specific examples on how to deal with categorical factors in regression analysis in practical world.

October 11
Time: 1:10–1:40 pm
Speaker: Shaofei Zhao
Title: Use Different Model to Predict the Temperature of Binghamton

Abstract: Weather forecasting has always been interesting and challenging, since we have already evaluated dataset aatemp during our short talk, using linear model may not get a good consequence for a variety of reasons, so it's necessary for us to consider making some changes both at the dataset and the methods, after these changes, we are supposed to get a better model and more precise predictions. This time we go to the same website U.S. Historical Climatology Network and download average daily temperature in Binghamton, and perform several different methods to fit the data, including linear model, time series model, etc. We will also test the goodness of fit and other model assumptions. Finally, we use different models to predict the future temperature, and make comments..

October 11
Time: 1:40–2:10 pm
Speaker: Yifei Zeng
Title: Application of Multidimensional Scaling

Abstract: Multidimensional Scaling (mds) is a method to reduce high-dimensional-data to low-dimensional-data, most likely to 2-dimensions or 3-dimensions which is the dimension that human can visualize. Also mds can also be used to cluster different groups of data. In this talk, we will discuss different types of mds and apply them to a data set from Izenman. The data set contains the distance between 48 cities in UK which forms a dissimilarity matrix. We will compare different types of mds with respect to visualization, computation, and fits of data etc. Also within a distinct type of mds, we will talk about how to choose the dimension so when we are reducing the dimension, we will still get enough information from data. We will finally present the map of the 48 cities and compare with the mds’ map to see the fit of mds.

October 18
Time: 1:10–1:40 pm
Speaker: Gang Cheng
Title: Problem with the error

Abstract: The Robust Regression method down weighting the extreme cases, but sometimes, when the large errors are sufficient numerous and extreme in value, it still failing. We need methods which fit the data well even in the presence of such problems. Least Trimmed Regression is a good way for dealing with data with many bad entries. Unlike Robust Regression, it gives no weights on such bed entries. In terms of determining the significance of variables, the R does not provide us with the standard error or p.value. Instead, we can solve it by bootstrap. In this talk, I will focus on the properties of LTS estimator and give examples of its working in real situation.

October 18
Time: 1:40–2:10 pm
Speaker: Schepis,Michael
Title: Classification Methodology Using Examples from Izenmana's Modern Multivariate Statistical Techniques

Abstract: One of the primary motivations of analyzing data is our ability to accurately categorize it. This is the purpose of classification; Different statistical techniques allow us to divide categorical data into relevant groups, however this ability is only as useful as our ability to interpret these groups in a meaningful way. Most data of interest cannot be perfectly separated without error, so it is also crucial to analyze misclassification rates with each of these statistical tools. In this talk, we will provide relevant examples from Alan J. Izenmanâ's \emph{Multivariate Statistical Techniques} and discuss the tools of Linear Discriminant analysis, Bayes Rule, and Quadratic Discriminant Analysis and when each of those techniques would be most appropriate and how to create a confusion matrix to examine the accuracy of our findings. For each technique we will discuss how the classifyier functions and give examples in which it might be applied.

October 25
Time: 1:10–1:40 pm
Speaker: McInroy,Alexander
Title: Binomial Regression and its Applications in Medical Diagnosis

Abstract: Regression analysis remains one of the most directly applicable tools in the field of statistics and machine learning today. In particular, binomial regression is especially useful for diagnosing medical conditions where a test leads to a positive or negative result. This talk will use examples from Extending the Linear Model with R, Faraway to illustrate this. Technical topics covered will be data analysis, data interpretation, model selection, and model prediction. Furthermore, selecting appropriate cutoff points for a binomial model’s response will be discussed, since mitigating type I and type II error is a subjective balancing act depending on the situation. Special attention will be given to selecting an appropriate cutoff point.

October 25
Time: 1:40–2:10 pm
Speaker: Rovou,Joshua
Title: Managing Multinomial Data: With examples from Dr. Ganggang Xu and Faraway's Extending the Linear Model

Abstract: Multinomial data requires careful thought to properly analyze. There are various forms that multinomial data can take that can easily be misclassified, leading to false conclusions. Understanding when, why, and how to recognize and apply these forms and their assumptions allows a data scientist to be mindful of best practices when fitting multinomial models. This paper provides an overview, suggestions, and examples of classifying multinomial data. In addition, this paper provides discussions on applying the appropriate assumptions and models in the R programming language, using the problem and data sets in Faraway’s “Generalized Linear Models” Chapter 5 as case studies. These examples seek to illustrate best practices when dealing with multinomial data for the student data scientist.

November 1
Time: 1:10–1:40 pm
Speaker: Xiaolin Tang
Title: MODEL SELECTION

Abstract: For a dataset, we can come up with plenty of models, however, there should be a best one through all of them. Therefore, we need some criteria to evaluate the model, and a method to find out it, here comes model selection.

My presentation will be presented with two parts: first, I will give a quick review of some basic concepts, including the criteria of a good model and the method to do model selection. Second, I will talk about several questions of chapter 10 in “Faraway I”(Linear Models with R, second Edition). To be specific, 10.1,10.4,10.5 and 10.6 will be covered. 10.1 is basic a review of model selection method. 10.4 talks about how different the reduced model can be with different selection direction, by direction I mean forward, backward etc. And the answer for this problem is the model will be quite different. 10.5 is about how outliers influence the result of model selection, we use the data with and without the outliers to do model selection separately, and we obtain the same model but with different estimated coefficient. And lastly, we will talk about how model selection infects prediction interval, it will be a little bit different from full model.

November 1
Time: 1:40–2:10 pm
Speaker: Xiang Wang
Title: Linear Discriminant Analysis for Diabetes Data

Abstract: Diabetes data resulted from a study conducted at the Stanford Clinical Research Center of the relationship between the three clinical classifications and five measurements for 145 instances. It helps the diagnosis and appropriate treatment to the diabetes patients. Based on all five variables representing the problematic multivariate Gaussian distributions and the inappropriate assumption of equal covariance matrices, misclassification rates in the leave-one-out cross-validation are 11% for LDA and 9.7% for QDA. Thus we will talk about the LDA via Multiple Regression and Logistic Discrimination, which is a semi-parametric model and asymptotically less efficient than is Gaussian LDA. In the view of Logistic Discrimination, we can use data transformation, variable selection techniques, polynomial regression, and other powerful methods to improve the result of LDA and QDA. Then we verify that when the Gaussian distributional assumptions or the common covariance matrix assumption are not satisfied, Logistic discrimination performs much better, and is more robust to non-normality than Gaussian LDA.

November 8
Time: 1:10–1:40 pm
Speaker: Wangshu Tu
Title: Parameters Selection in Guassian Kernel SVM

Abstract: In Support Vector Machines(SVM), to get small misclassification rate(MCR), we need to adjust C, which is regularization parameter to constrain the range of Lagrangian coeffcients $\alpha$ in dual function FD, and parameters $\tau$ in kernel function. But often it is not so clear that how to select them. In this long talk, Gaussian radial basis function(RBF) will be used as kernel function, because only one parameter(C,$\tau$ ) needs to be determined. In order to get desired parameter, one method is to find minimal missclassification error by using 10-folder Cross Validation(CV/10) with different pairs of (C, $\tau$ ). Based on that, other optimization methods, and An Automatic Method for Selecting the Kernel Parameter $\tau$ will be discussed.

November 8
Time: 1:40–2:10 pm
Speaker: Yangwei Jiang
Title: Lineal Model diagnostics

Abstract: When a linear model has been constructed, we need to check whether the model is appropriate and do some adjustment if necessary, this process is called model diagnostics. There are several aspects that we need to consider, during the short talk, we have discussed about the error term. In this long talk, we will discuss about the outliers, leverage and influential points, and we will also introduce partial regression plot and partial residual plot to detect outliers and non-linearity of the regression model.

Spring 2017

February 23
Time: 1:15–2:15 pm
Speaker: Yu Hu
Title: Vehicle's Fuel Economy Data Analysis

Abstract: This project I did aimed at providing reliable estimates for comparing vehicles in 2016. The purpose is to help car buyers choose the most fuel-efficient vehicle that meets their needs. Using the linear regression knowledge to analysis the correlation between each factor(vehicles's weight, cylinder, engine size, etc.) and response(fuel economy).
March 2
Time: 1:15–2:15 pm
Speaker: Hao Wang
Title: On testing independence and goodness of fit in linear models

Abstract: We consider a linear regression model and propose an omnibus test to simultaneously check the assumption of independence between the error and the predictor variables and the goodness of fit of the linear regression model.
March 16
Time: 1:15–2:15 pm
Speaker: Liping Gu
Title: Analysis of the Dataset “Wine”

Abstract:In the talk, I discuss the statistical analysis on the dataset “wine”. First I converted the numerical predictors into factorial ones and used the method which is similar to the factorial design to find the impact of different factors.Then making use of the algorithms of LDA and QDA I analyzed the data. The result of the computation shows a clear tendency on the means of the predictors that stand out as significant in the result of the factorial model. The data analysis suggests that 4 predictors are significant, whereas the other 7 are not.

May 2
Time: 12–1 pm
Speaker: Hao Wang
Title: On testing independence and goodness of fit in linear models

Abstract: We consider a linear regression model and propose an omnibus test to simultaneously check the assumption of independence between the error and the predictor variables and the goodness of fit of the linear regression model.

Statistics Seminar

2026-05-04T16:27:00-04:00

Statistics Seminar

The Statistics seminar aims to cover topics from all areas of statistics both from a traditional perspective but also from a more data science perspective. The seminar is also offered as MATH 567, Seminar in Statistics, Section 01.

Location: Whitney 100E (See the directions to the department)
Time: Thursdays, from 1:30 pm to 2:30 pm
Organizers: Jingchen (Monika) Hu and Pratik Misra

See also: Data Science Seminar and the Capstone Seminar.

Spring 2026

January 29
Speaker: Bahareh Baharinezhad (Binghamton University)
Title: A case study in option trading
Abstract

February 12
Speaker: Nils Sturma (EPFL)
Title: Matching criterion for identifiability in sparse factor analysis
Abstract

February 26
Speaker: Andrés Felipe Barrientos (Florida State University)
Title: Bayesian nonparametric modeling of mixed-type bounded data
Abstract

March 5
Speaker: Aliu Adebiyi (Binghamton University)
Title: Bayesian hierarchical pathway-structured model for RNA-seq differential expression
Abstract

March 12
Speaker: Samruddhi Abhay Thakar (Binghamton University)
Title: Adaptive functional principal components analysis
Abstract

March 19
Speaker: Ruobin Gong (Rutgers University)
Title: Privacy differentials in differential privacy
Abstract

March 24
Speaker: Zifan Huang (Binghamton University)
Title: From roots to strict zero crossings: developing a well-posed definition of the Buckley-James estimator
Abstract

March 26
Speaker: Roman Semenko (Binghamton University)
Title: Reasoning under uncertainty: Dempster–Shafer theory and algorithmic approaches to belief updating
Abstract

April 9
Speaker: Yiyi Cao (Binghamton University)
Title: Bridging empirical auditing and theoretical privacy guarantees - toward robust and interpretable privacy evaluation in modern machine learning
Abstract

April 10
Speaker: Alex Markham (University of Copenhagen)
Title: Coarsening causal DAG models
Abstract

April 16
Speaker: Zhongyuan Zhao (Binghamton University)
Title: On the quasi-stationary distribution of the Shiryaev–Roberts recurrence driven by exponential data
Abstract

April 23
Speaker: Yangsheng Wang (Binghamton University)
Title: Functional autoencoder for smoothing and representation learning
Abstract

April 30
Speaker: Maryclare Griffin (UMass Amherst)
Title: Not there yet: practical stopping criteria for MCMC with many chains
Abstract

May 7
Speaker: Geran Zhao (Binghamton University)
Title: Deep canonical correlation analysis
Abstract

Fall 2025

August 28
Speaker: Pratik Misra (Binghamton University)
Title: Structural identifiability and causal discovery in Gaussian graphical models
Abstract

September 18
Speaker: Zehang Richard Li (UC Santa Cruz)
Title: Robust cause-of-death assignment using verbal autopsies
Abstract

September 25
Speaker: Bryon Aragam (University of Chicago)
Title: Bridging causality and deep learning with causal generative models
Abstract

October 9
Speaker: Katherine Grzesik (University of Rochester)
Title: Adapting graduate-level statistics coursework for the undergraduate statistics major
Abstract

October 16
Speaker: Samuel Wang (Cornell University)
Title: Confidence sets for causal orderings
Abstract

October 30
Speaker: Mathias Drton (Technical University of Munich)
Title: Causal modeling with stationary processes
Abstract

November 6
Speaker: Harrison Quick (University of Minnesota)
Title: Reliable rates in disease mapping
Abstract

November 13
Speaker: David Collins (Binghamton University)
Title: Bayesian D-optimal design of experiments with quantitative and qualitative responses
Abstract

November 20
Speaker: Bruce Phillips (Binghamton University)
Title: Subdata selection for Principal Component Analysis
Abstract

December 4
Speaker: Sanvesh Srivastava (University of Iowa)
Title: Bayesian compressed mixed-effects models
Abstract

Spring 2025

March 6
Speaker: Yushu Shi (Weill Cornell Medicine)
Title: CAT: A conditional association test for microbiome data using a permutation approach
Abstract
March 20
Speaker: Praveen Niranda (Internal)
Title: Network Reconstruction Using Nonparametric Additive ODE Models
Abstract
March 27
Speaker: Jonathan Stallrich (NC State)
Title: Optimal Designs for Two-Stage Inference
Abstract
April 3
Zoom presentation only
Speaker: Rumi Chunara (NYU)
Title: A Multi-level Perspective for Navigating the Intersection of Data and Public Health
Abstract
April 10
Speaker: Tanzy Love (University of Rochester)
Title: TBD
Abstract
April 17
ABD Exam presentation
Speaker: Zhongyuan Zhao (Internal)
Title: On Optimality of the Shiryaev-Roberts Change-Point Detection Method in the Exponential Case
Abstract
April 24
Speaker: Bruce Phillips (Internal)
Title: Data Twinning
Abstract
May 1
Speaker: Xinhai Zhang (Internal)
Title: Neural Network Models in CATE Estimation
Abstract

Geometry and Topology Seminar

2026-04-19T08:37:30-04:00

Geometry and Topology Seminar

We meet Thursdays at 2:45–3:45 pm in Whitney Hall 100E. This semester's organizers are James Hyde and Lorenzo Ruffoni. The seminar has an announcement mailing list open to all.

Topics include: geometric group theory, differential geometry and topology, low-dimensional topology, algebraic topology, and homotopy theory.

Spring 2026

January 29h
Speaker: Juliet Aygun (Cornell)
Title: Counting geodesics on prime-order k-differentials

Abstract: It has been of popular interest over the last several decades to count geodesics with respect to their length on flat surfaces. Asymptotics of these counting functions for generic translation surfaces, which are Riemann surfaces with a holomorphic one form, have been determined by the pioneering work of Eskin-Masur-Zorich. There is a more general type of flat surface called a (1/k)-translation surface, which is a Riemann surface with a k-differential. Equivalently, a (1/k)-translation surface is a collection of polygons in the complex plane with sides identified pairwise by translation and possible rotations of 2pi/k. In this talk, we will discuss asymptotics of these counting functions on generic (1/k)-translation surfaces when k is prime and genus is more than two. The main tools I will discuss are GL+(2,R)-orbit closures and a result of Eskin-Mirzakhani-Mohammadi which relates asymptotics to GL+(2,R)-orbit closures.

February 5th
Speaker: Barry Minemyer (Commonwealth University - Bloomsburg)
Title: Negatively Curved Metrics on Complex Hyperbolic Branched Covers

Abstract: Gromov and Thurston used hyperbolic branched cover manifolds to construct the first known examples of compact manifolds which admit a pinched negatively curved metric, but do not admit a hyperbolic metric. Fine and Premoselli (n=4) and Hamenstadt and Jackel (n > 4) later used these same manifolds to construct the first known examples of negatively curved Einstein metrics (in these respective dimensions) that are not locally symmetric.

Recently, Stover and Toledo proved that analogous complex hyperbolic branched cover manifolds exist. They also proved that these manifolds do not admit a locally symmetric metric, and a result of Zheng shows that these manifolds are Kahler. In this talk I will present recent work proving the existence of pinched negatively curved metrics, as well as the existence of negatively curved Kahler-Einstein metrics (due to Guenancia and Hamenstadt) on these complex hyperbolic branched cover manifolds. Part of my presented work is joint with Lafont.

February 12th
Speaker: Francis Wagner (Cornell University)
Title: Isoperimetric functions and the Word Problem

Abstract: A fundamental algorithmic question in group theory is the Word Problem for finitely generated groups, which asks whether there exists an algorithm to decide whether two words on the generators represent the same group element. A related notion is the Dehn function of a finitely presented group, the smallest isoperimetric function of the presentation's Cayley complex. While the Dehn function gives an upper bound for the complexity of the Word Problem for that group, this bound is only meaningful in the class of finitely presented groups and is very far from sharp even in this class. We resolve this disconnect by instead considering the Dehn functions of the finitely presented groups into which a group embeds, demonstrating a refinement of the Higman embedding theorem that gives a potentially quasi-optimal bound on the Dehn function of the ambient group.

February 19th
Speaker: Satya Howladar (University of Florida)
Title: Gromov’s Conjecture for Product of Baumslag-Solitar Groups and some other One relator groups

Abstract: Gromov introduced macroscopic dimension of metric spaces in order to study large scale properties of manifolds. He conjectured that a closed $n$-manifold which admits Positive Scaler Curvature metric, should have its universal cover to be of macroscopic dimension at most $n-2$, with respect to the pull back metric on it. This conjecture depends a lot on the fundamental group of the base manifold. For $n>4$, closed spin $n$-manifolds $M$, we developed sufficient condition on $\pi_1(M)$, to verify the conjecture. When $\pi_1(M)$ is product of $2$-dimensional groups (i.e. groups with classifying space a $2$-dimensional CW complex), $\mathbb Z_2$-summands in their homology creates a problem for application of our technique. We could resolve this in the case of certain one-relator groups, including Baumslag-Solitar, and certain others, by passing to some finite index subgroup of them not admitting $\mathbb Z_2$-torsion in homology. This is done by the well-known technique of Fox calculus, to analyze boundary maps of cells of finite index covers. I will try to revisit this technique and sketch a proof our result.

February 26th
Speaker: Lucas Williams (Binghamton University)
Title: Equivariant Framed 1-Manifolds and the Pontryagin-Thom Isomorphism

Abstract: The Pontryagin-Thom construction gives an isomorphism between the cobordism group of framed n-manifolds and the nth stable homotopy group of the sphere spectrum. The G-equivariant Pontryagin-Thom construction gives an isomorphism between the cobordism group of V-framed G-manifolds and the Vth stable homotopy group of the G-equivariant sphere spectrum. We will discuss both of these constructions and then present some new explicit descriptions of the images of each 1-dimensional manifold equipped with an action by the cyclic group of order 2 in their relevant homotopy groups. We subsequently provide a new perspective on some key differences between the equivariant and non-equivariant Hopf fibration.

March 5th
Speaker: Oliver Wang (University of Virginia)
Title: Group actions on exotic spheres

Abstract: In the 1960's W.C. Hsiang and W.Y. Hsiang showed that exotic spheres admit less symmetries than the standard sphere. However, constructing symmetries on exotic spheres has been a difficult task. It is still an open question whether or not every exotic sphere admits a smooth, nontrivial S^1-action. In fact, it is open whether or not every exotic sphere admits a smooth, nontrivial C_p-action, where C_p denotes the cyclic group of order p. In this talk, I will discuss recent work with Nick Kuhn and J.D. Quigley relating this problem to stable homotopy theory.

March 12th
Speaker: Sofia Martinez (Bryn Mawr College)
Title: Localizations on Lattices

Abstract: Homotopical Combinatorics is a newer area of Algebraic Topology that studies in a more tractable manner the homotopical structure of topological spaces with the action of a group. The main objects of study in this new area are Transfer systems, originally created with the goal of understanding equivariant analogs of higher coherences. In a more category-theoretic language, a transfer system on a poset, or more generally a finite category, C, is a wide subcategory of C closed under pullbacks. This talk will focus on the case when C=Sub(G), the subgroup lattice of a finite group, G. Subsequent work shows that transfer systems occur naturally as the acyclic fibrations of nicer categories known as model category and their are universal constructions that provide a way to move between these categories; these are called left and right Bousfield localizations. In this we will see how transfer systems change under these types of constructions.

March 13th
PETER HILTON MEMORIAL LECTURE
SPECIAL TIME AND LOCATION: March 13, 3:30pm, Alumni Lounge at Old O'Connor Hall
Speaker: Martin Bridson (University of Oxford)
Title: Chasing finite shadows of infinite groups through geometry

Abstract: There are many situations in geometry or elsewhere in mathematics where it is natural or convenient to explore infinite groups of symmetries via their actions on finite objects. But how hard is it find these finite manifestations and to what extent does the collection of all such actions determine the infinite group?

In this colloquium, I will sketch some of the rich history of such problems and then describe some of the great advances in recent years. I'll describe pairs of distinct groups that have the same finite images and I'll sketch the proof of some “profinite rigidity results”, i.e. theorems showing that in certain circumstances one can identify an infinite group if one knows its set of finite images.

March 19th
Speaker: Francesco Lin (Columbia)
Title: Coexact 1-form spectral gaps of hyperbolic rational homology spheres

Abstract: The spectral gap of the Hodge Laplacian of functions (or, equivalently, exact 1-forms) is a very well-studied fundamental quantity associated to a hyperbolic three-manifold. In recent years, the problem of understanding its counterpart on coexact 1-forms has also spurred a lot of activity because of its relation with questions in number theory and low-dimensional topology. In this talk, after introducing the geometric setup and highlighting some fundamental differences between these two quantities, I will focus on some structural properties of the set of coexact 1-form spectral gaps of hyperbolic rational homology spheres. In particular, I will discuss a construction that allows to determine somewhat explicitly some interesting accumulation points of the set of such spectral gaps. This is joint work with M. Lipnowski.

March 26th
Speaker: Varinderjit Mann (Cornell University)
Title: Slice enriched categories, microcosm principle, and slice model structures

Abstract: The goal of this talk is to highlight an example of a very important phenomenon known as the microcosm principle, coined by John Baez and James Dolan. This particular example involves discussing pseudomonoid objects and pseudomodule objects. Furthermore, the discussion is interesting in its own right and allows one to define general sliced enriched categories. We will begin by reviewing what it means to categorify general algebraic structures such as monoids. Then, we will take this further to the general context of horizontal categorification and vertical categorification with examples. Our primary consideration of this will be that of a pseudomonoid object and pseudomodule object in a monoidal 2-category. Finally, we discuss the slicing of pseudomonoid objects and pseudomodule objects in the monoidal 2-category Cat, and describe how they respect the consideration of model structures.

April 2nd

(Spring break - no seminar)

April 9th
Speaker: Sanjana Agarwal (Indiana University, Bloomington)
Title: Dennis trace for combinatorial K-theories

Abstract: Classically, algebraic K-theory captures various invariants associated to a ring R used widely in algebraic and arithmetic geometry and number theory. To compute these invariants, one of the most successful tools have been trace methods. The trace method machinery builds off a map called the Dennis trace map from the algebraic K-theory of R to the Hochschild homology of R. In recent years, new analogues of algebraic K-theory have been introduced (first by Zakharevich) in `combinatorial' categories motivated by generalized Hilbert's third problem. In this talk, we present initial attempts to generalize the theory of trace methods to such combinatorial K-theories. This is joint work with Ramyak Bilas.

April 16th
Speaker: Josefien Kuijper (University of Toronto)
Title: All K-theory is squares K-theory: constructing a derived Euler characteristic

Abstract: Combinatorial (or “cut-and-paste”) K-theory is a modern approach to the study of the classical polytope scissors congruence groups, inspired by algebraic K-theory of Waldhausen categories, and can be applied to other geometric settings as well, such as the categories of varieties and semi-algebraic sets. We present the K-theory of squares category as a framework that unifies Waldhausen K-theory as well as many instances of combinatorial K-theory. As an application, we lift the Euler characteristic for definable sets in an o-minimal structure to a map of K-theory spectra.

April 23th
Speaker: Changjie Chen (CRM Montreal)
Title: Morse theory on moduli of curves and stability phenomena

Abstract: In 1997, Sarnak conjectured that the determinant of the Laplacian is a Morse function on the space of unit area Riemannian metrics on a given real surface, and hence induces a Morse function on its moduli space. Meanwhile, the systole function, defined as the length of a shortest essential closed geodesic with respect to the base Riemannian metric, is topologically Morse on the Teichmüller space of n-dimensional flat tori (due to Ash) and of Riemann surfaces of genus g with n marked points (due to Akrout), though it does not yield a classical Morse theory. In this talk, I will introduce a family of Morse functions, defined as Gaussian averages of the length spectrum, on the Deligne–Mumford compactification (M_{g,n} bar). These functions are compatible with the Deligne–Mumford stratification and the Weil–Petersson metric, and their critical points can be characterized by a combinatorial property. I will talk about the index gap theorem for these functions and its homological consequences, in the form of a stability theorem for the homology of moduli spaces of stable curves. If time permits, I will also explain how these Morse functions connect to Sarnak’s conjecture.

April 30th
Speaker: Urshita Pal (University of Michigan)
Title: The top cohomology of principal congruence subgroups of $SL_nR$ and $Sp_{2n}R$

Abstract: I will discuss the rational cohomology of $SL_nR$, $Sp_{2n}R$, and their principal congruence subgroups, for R a number ring. Borel–Serre showed that these groups satisfy a (co)homological duality that lets us study their cohomology groups via certain representations called `Steinberg modules’, which have a combinatorial description in terms of Tits buildings. I will describe forthcoming work that uses this approach to compute the top cohomology of certain principal congruence subgroups of prime level, when R is a Euclidean domain. The computations rely on the structure of the units in the field obtained by quotienting R by the prime, modulo the units coming from R itself.

Hosting Seminar Speakers: Guidelines and Procedures

2024-11-19T15:48:04-04:00

Hosting Seminar Speakers: Guidelines and Procedures

If you are hosting a visiting speaker, here are the key procedures to follow regarding arrangements, reimbursements, and departmental support.

1. Initial Arrangements

Invitation Confirmation: As the host, you must forward to mathoffice@binghamton.edu the email trail showing the invitation and acceptance correspondence, including the seminar date.
Itinerary: If the speaker will be staying for more than one day (e.g., arriving the night before the seminar), provide a itinerary outlining the daily schedule, including times and activities.

2. Reimbursement Procedures

Expense Submission: If your speaker will be submitting expenses for reimbursement, they must visit the Department Office before 4 PM on the last of day of the visit to complete the necessary reimbursement forms.

3. Logistical Support

Hotel Reservations: Email mathoffice@binghamton.edu with the speaker's name and their check-in/check-out dates.
Visitor Office: If an office is needed for the day, notify the administrative assistants who will arrange an office and key.
Parking Passes: Speakers can pick up a parking pass from the Department Office. Visitor parking is available in the parking garage or the lot across from Bartle Library.

4. Dining Arrangements

BU Dining Services:
- Obtain a CrediDine card from the administrative assistants before dining.
- Return the card and original itemized receipt the same day.
- The department covers meals for the speaker, host faculty and one more faculty member.
Dining at a Restaurant:
- Submit the original itemized receipt showing the payment method to the administrative assistants.
- Alcohol and tax are not reimbursed (a tax-exempt form is available, though not all places accept it).
- A tip up to 20% is reimbursable.
- The department covers meals for the speaker, host faculty and one more faculty member.
Maximum allowable expense per person: Lunch = 25.00 (incl. tip), Dinner = 45.00 (incl. tip).

5. Travel Arrangements

Public Transportation (Bus/Uber/Lyft, etc.):
- Submit the original itemized receipt showing the payment method for reimbursement.
- Electronic receipts can be emailed to mathoffice@binghamton.edu, while paper receipts should be submitted by email, mail or in person. The original paper form or a scanned copy is needed for reimbursement.
Car Travel:
- Rental Car: NY State reimburses rental car expenses plus gas (with original receipts). The rental must be compact or standard size. Visitors must book their own rental; do not offer the NYS/BU account number. The base rate and damage waiver will be reimbursed (for compact or standard size). NYS will not reimburse liability insurance.
- Personal Car: Mileage is reimbursed based on the distance from the visitor's address to Binghamton University, following the NYS/IRS standard mileage rate. Gas receipts will not be reimbursed, as fuel costs are included in the mileage rate. New York State will reimburse the lower amount between the mileage costs (calculated using NYS/IRS rates) and the market rate for rental cars. Please ensure your visitor is informed of these reimbursement limits.
Air Travel:
- Please consult with Dianne for the latest guidelines regarding visitor air travel, as these rules frequently change.