Tools

Software released by the group — research code, model checkers, simulators, and analysis pipelines, distributed for re-use by the community. Each tool links back to the publication that introduced its methodology.

• PGTGProcedurally Generated Grid-Based Traffic Gym is a Gymnasium-compliant, modular reinforcement learning benchmark for studying exploration, safety, and generalization in a single, controllable environment.
• PyDSMCStatistical Model Checking for Neural Agents Using the Gymnasium Interface. An open-source Python library for easy statistical model checking and accurate reporting of agents conforming to the Gymnasium API.
• SepDistancesSeparation distances and s/c-metrics for comparing causal graphs: DAGs, CPDAGs, mixed and cyclic graphs, with SHD and AID for reference.