New preprint: differentiable discrete abstractions

Our team figured out how to optimize discrete abstractions of continuous-state systems with gradient descent. The solution lies in a smooth proxy for a simulation metric. An example state grid, resulting from our technique, is shown above.

Details below:

• Jordan Peper, James Mathias Gast, Vignesh Nanduri, Tanmayee Maram, Ethan Howes, Ivan Ruchkin.
  S3: A Smooth Simulation Surrogate for Optimizing Discrete Abstractions of Dynamical Systems [Arxiv] [Github].
  Preprint, 2026.