Georgia Institute of Technology College of Sciences

The first meeting of our new professional development seminar for postdocs and other interested individuals (such as advanced graduate students). A discussion of the triumvirate of faculty positions: research, teaching, and service.

Dirac'stheory of constrained Hamiltonian systems allows for reductions of the dynamics in a Hamiltonian framework. Starting from an appropriate set of constraints on the dynamics, Dirac'stheory provides a bracket for the reduced dynamics. After a brief introduction of Dirac'stheory, I will illustrate the approach on ideal magnetohydrodynamics and Vlasov-Maxwell equations. Finally I will discuss the conditions under which the Dirac bracket can be constructed and is a Poisson bracket.

In 2011, Rothvoß showed that there exists a 0/1 polytope such that any higher-dimensional polytope projecting to it must have a subexponential number of facets, i.e., its linear extension complexity is subexponential. The question as to whether there exists a 0/1 polytope having high PSD extension complexity was left open, i.e. is there a 0/1 polytope such that any spectrahedron projecting to it must be the intersection of a subexponential sized semidefinite cone and an affine space? We answer this question in the affirmative using a new technique to rescale semidefinite factorizations

Maximum likelihood estimation is a fundamental computational task in statistics and it also involves some beautiful mathematics. The MLE problem can be formulated as a system of polynomial equations whose number of solutions depends on data and the statistical model. For generic choices of data, the number of solutions is the ML-degree of the statistical model. But for data with zeros, the number of solutions can be different. In this talk we will introduce the MLE problem, give examples, and show how our work has applications with ML-duality.This is a current research project with Elizabeth Gross.

Equilibrium computation is among the most significant additions to the theory of algorithms and computational complexity in the last decade - it has its own character, quite distinct from the computability of optimization problems. Our contribution to this evolving theory can be summarized in the following sentence: Natural equilibrium computation problems tend to exhibit striking dichotomies. The dichotomy for Nash equilibrium, showing a qualitative difference between 2-Nash and k- Nash for k > 2, has been known for some time. We establish a dichotomy for market equilibrium. For this purpose. we need to define the notion of Leontief-free functions which help capture the joint utility of a bundle of goods that are substitutes, e.g., bread and bagels. We note that when goods are complements, e.g., bread and butter, the classical Leontief function does a splendid job. Surprisingly enough, for the former case, utility functions had been defined only for special cases in economics, e.g., CES utility function. We were led to our notion from the high vantage point provided by an algorithmic approach to market equilibria. Note: Joint work with Jugal Garg and Ruta Mehta.

Abstract: We develop techniques for the verication of the Chebyshev property of Abelian integrals. These techniques are a combination of theoretical results, analysis of asymptotic behavior of Wronskians, and rigorous computations based on interval arithmetic. We apply this approach to tackle a conjecture formulated by Dumortier and Roussarie in [Birth of canard cycles, Discrete Contin. Dyn. Syst. 2 (2009), 723781], which we are able to prove for q <= 2.