Shanshan Ding (Google)
If you want to join the room-sharing list, please fill out this form.
Pod 39 (13 mins walk): use the promo code “podpal” on their website
Alex Fink: Speyer's tropical f-vector conjecture and its proof
In 2008, looking to bound the face vectors of tropical linear spaces, Speyer introduced the g-invariant of a matroid. He proved its coefficients nonnegative for matroids representable in characteristic zero and conjectured this in general. After introducing these objects, this talk will give an overview of the ingredients in recent work with Andrew Berget that proves the conjecture. A main character is the variety of coordinatewise quotients of points in two linear subspaces, and its initial degenerations which encode a new generalization of external activity to a pair of matroids.
Emily Gunawan: Pattern-avoiding c-Birkhoff polytopes
There is a bijection between type A Coxeter elements c and type A Dynkin quivers Q. For each type A Coxeter element c, we define a pattern-avoiding Birkhoff subpolytope whose vertices are the permutation matrices of the c-singletons. We show that the (normalized) volume of our polytope is equal to the number of longest chains in a corresponding type A Cambrian lattice. Our work extends a result of Davis and Sagan which states that the volume of the convex hull of the 132 and 312 avoiding permutation matrices is the number of longest chains in the Tamari lattice, a special case of a type A Cambrian lattice. Furthermore, we prove that each of our polytopes is unimodularly equivalent to the (Stanley’s) order polytope of the heap H of the longest c-sorting word. The Hasse diagram of H is given by the Auslander—Reiten quiver for the quiver representations of Q. This talk is based on joint work with Esther Banaian, Sunita Chepuri, and Jianping Pan.
Eric Ramos: Universality theorems for generalized splines
We study generalized splines from the perspective of the representation theory of the category of graphs with contractions. We prove a finite generation result in this context, which implies a kind of universality in the generating set for the module of splines over any graph within a fixed homotopy class. Importantly, these results hold over any Noetherian commutative ring with a chosen finite list of ideals for edge-labels. One can think of our results as expanding upon and explaining many results about generating sets of the module of generalized splines.
Eric Rowland: Combinatorial structure behind Sinkhorn limits
The Sinkhorn limit of a positive square matrix is obtained by scaling the rows so each row sum is 1, then scaling the columns so each column sum is 1, then scaling the rows again, then the columns again, and so on. It turns out that the sequence of matrices obtained in this way converges, and the limit has been used for almost 90 years in applications ranging from predicting telephone traffic to machine learning. But until recently, nothing was known about the exact values of its entries. In 2020, Nathanson determined the Sinkhorn limit of a 2 × 2 matrix. Shortly after that, Ekhad and Zeilberger used Gröbner bases to determine the Sinkhorn limit of a symmetric 3 × 3 matrix. Recently, Jason Wu and I determined the Sinkhorn limit of a general 3 × 3 matrix. This was enough to guess the general form for n × n matrices. However, the values of the coefficients that arise are not obvious. We used 1.5 years of CPU time to interpolate formulas from numeric examples, leading to a conjectural determinant formula that reflects new combinatorial structure on sets of minor specifications.