Coxeter groups are abstract reflection groups. They come up in the study of symmetries of polytopes, but there are also affine and hyperbolic types of symmetry groups as well. In my own work, I primarily study finite Coxeter groups from a combinatorial viewpoint. The paradigm here is that one begins with a combinatorial property for permutations and interprets it as an algebraic or geometric property. For example, the number of inversions of a permutation is its minimal length when expressed as a word in terms of simple transpositions.
Combinatorics has been called the art and science of counting. When most people think of combinatorics, they are thinking of enumerative combinatorics, which studies and classifies properties of combinatorial sets, like permutations, lattice paths, or partitions. The goal is to find good exact formulas when possible, often expressed in terms of tools known as generating functions. Other times, the task is to explain why two seemingly different formulas enumerate the same set of objects. Some immediate consequences of enumerative combinatorics come in the form of probabilistic results. For example, the tools of enumerative combinatorics are used heavily in the study of mixing times for card shuffling, as discussed in the work of Diaconis and others. I have done a lot of work in pure enumerative combinatorics, mainly with respect to permutation statistics.
Combinatorial models can be used as a way to approximate a continuous situation with a very large but finite set of possibilities. A random combinatorial object from the finite set can then serve as a proxy for the typical state of the continuous system. If we're lucky, the analysis of the discrete model yields new insights. One big result of this flavor is the work of Vershik-Kerov and Logan-Shepp, who found that Young diagrams under Plancharel measure have an expected limit shape when appropriately rescaled. As a consequence, they answered a question of Ulam to affirm that the longest increasing subsequence of a random permutation of \(n\) numbers is about \(2\sqrt{n}\). My own work in this area has been fairly modest, but I have studied large random models in a couple of projects, and thoughts or randomness and randomization have always fascinated me.
Topology seems very far from combinatorics in spirit, but really combinatorics is at the heart of much of topology from its roots in Euler's formula: V-E+F = 2. There are vast generalizations of this result, and the open problems in this field often boil down to the combinatorics of abstract faces under certain constraints on how they can attach to one another. I have contributed a couple of ideas to this field, especially in my collaboration with Nevo, and with Nevo and Tenner.
Who doesn't love the study of the integers? I am a dabbler, really, but there have been one or two times that I was able to say something interesting about properties of whole numbers. My work on Sos permutations partially falls into this category. Currently, I am exploring properties of "generalized triangular numbers" with Validashti.
Affine Descents and the Steinberg torus (2007 AMS meeting)
What does the gamma vector count? (2010 AMS meeting)
A funny thing happened on the way to Steppenwolf Theatre... (2012 MAA MathFest)
Combinatorics of the shard intersection order (2013 AMS Meeting)
The module of affine descent classes (2013 FPSAC conference)
Depth of an element of a Coxeter group (2014 AMS meeting)
Reflection length in affine Coxeter groups (2018 FPSAC conference)
Farey Permutations (2019 AMS meeting)
Zig-zag Eulerian polynomials: from A_n(t) to Z_n(t) (2024 AMS meeting)