Joseph Scott
AI, automated reasoning, and machine learning
I am a doctoral student (PhD) at the University of Waterloo in the David Cheriton School of Computer Science, where I am tremendously lucky to be supervised by Vijay Ganesh. Previously, I received my masters (M.Math) in computer science from the University of Waterloo and a bachelors in Mathematics (B.S) and Computer Science (B.S) from Ohio University.
My current research focus is on Artificial Intelligence (AI) and its applications. I am particularly interested in engineering systems that dually leverage deductive reasoning tools with abductive and inductive reasoning tools. Specifically, my research focuses on:
- Automated Reasoning: propositional satisfiability; satisfiability modulo theories; mixed-integer linear programming; and solvers
- Machine Learning: meta-learning; algorithm selection; active learning; machine teaching; symbolic regression
- AI Safety: verification of neural networks
- Applications of AI: reinforcement learning driven fuzzing and testing; physics and automating scientific discovery; finance
Selected publications
- Algorithm Selection for SMT International Journal on Software Tools for Technology Transfer 25(2), 2023
- Goose: A Meta-Solver for Deep Neural Network Verification SMT Workshop at IJCAR/FLoC 2022, 2022
- BanditFuzz: Fuzzing SMT Solvers with Multi-agent Reinforcement Learning FM 2021, 2021
- MachSMT: A Machine Learning-based Algorithm Selector for SMT Solvers TACAS 2021, 2021
- BanditFuzz: A Reinforcement-Learning Based Performance Fuzzer for SMT Solvers VSTTE 2020, 2020