PhD Seminar • Artificial Intelligence • ML-Driven Algorithm Selection and Meta-Solving for Hard Combinatorial Problems

Please note: This PhD seminar will take place online.

Joseph Scott, PhD candidate
David R. Cheriton School of Computer Science

Supervisor: Professor Vijay Ganesh

On hard search problems, no single algorithm is expected to be the fastest on all inputs. For example, in practice, SAT/SMT solvers deploy several strategies ranging from algorithm selection to solver portfolios. In this talk, I’ll present two tools, MachSMT — an algorithm selection tool for Satisfiability Modulo Theories (SMT) solvers, and Goose — an adaptive algorithm selection tool, which we dub a meta-solver, for deep neural network verification.