Papers
10
Total Citations
73
H-Index
4
About
Stanley Bak is a researcher whose work sits at the intersection of formal verification, control systems, and autonomous cyber-physical systems (CPS), with a particular focus on making safety-critical systems provably safe. His research has made significant contributions to reachability analysis, runtime assurance, and the application of advanced mathematical tools — most notably polynomial zonotopes — to guarantee safety in complex nonlinear systems. Among his most impactful contributions is his work on provably safe reinforcement learning through action projection using reachability analysis and polynomial zonotopes (34 citations), which addresses one of the field's central challenges: enabling reinforcement learning agents to operate in safety-critical environments without sacrificing formal guarantees. His development and advocacy of the Simplex Architecture for runtime assurance of autonomous CPS has further established practical frameworks for deploying unverified controllers safely. Bak has also contributed foundational software tools for hybrid systems analysis, including C2E2, HyST, and TuLiP, broadening accessibility to verification methods for the research community. His ongoing investigations into sparse polynomial zonotope intersection checking and Koopman operator linearization reflect a commitment to both theoretical rigor and computational efficiency, making him a prominent voice in the formal methods and autonomous systems communities.
Research Focus
Key Achievements
Top Papers
- 1
- 2The Black-Box Simplex Architecture for Runtime Assurance of Autonomous CPS11 citations · 2022
- 3Conformant Synthesis for Koopman Operator Linearized Control Systems7 citations · 2022
- 4
- 5
- 6Safe CPS from unsafe controllers3 citations · 2021
- 7On the Difficulty of Intersection Checking with Polynomial Zonotopes3 citations · 2023
- 8Sparse Intersection Checking for Sparse Polynomial Zonotopes2 citations · 2025
- 9
- 10