John Wilmes

Assistant Professor of Mathematics

Departments/Programs

Mathematics
Volen National Center for Complex Systems

Degrees

University of Chicago, Ph.D.
University of Chicago, M.S.
Reed College, B.A.

Expertise

Combinatorics. Algorithms. Machine Learning.

Profile

My research is in the theory of computing and combinatorics. A major area of focus is at the
intersection of the algorithmic Graph Isomorphism problem and algebraic combinatorics. The analysis
of graph isomorphism algorithms is often closely related to the complexity of the automorphism
group, and I am therefore particularly interested in the connection between structure and symmetry
of graphs. Another primary area of interest is in computational learning theory, particularly in
provable guarantees for neural network algorithms.

Courses Taught

MATH	15a	Applied Linear Algebra
MATH	16b	Applied Linear Algebra Practicum
MATH	122a	Numerical Methods and Big Data
MATH	124a	Optimization

Scholarship

Santosh Vempala and John Wilmes. Gradient Descent for One-Hidden-Layer Neural Networks: Polynomial Convergence and SQ Lower Bounds. Proc. of Conference on Learning Theory. Phoenix, Arizona: Proceedings of Machine Learning Research, 2019.

Daniel Štefankovič, Eric Vigoda, and John Wilmes. On Counting Perfect Matchings in General Graphs. Proc. of Latin American Symposium on Theoretical Informatics. Buenos Aires, Argentina: Springer, 2018.