Please follow the readme instructions to construct experiments.

An illustration of continuous RL training pipeline for safety-critical black-box systems.

Detailed instructions on how to conduct experiments can be found in the readme. The following is a brief summary of the steps.

• Grid-Search Validation

cd ./algorithms
python grid_search.py

• Monte-Carlo-Search Validation

cd ./algorithms
python monte_carlo_search.py

• Cross-Entropy-Search Validation

cd ./algorithms
python ce_search.py

• Bayesian-Optimization-Search Validation

cd ./algorithms
python bayesian_search.py

• Adaptive-Multilevel-Splitting-Search Validation

cd ./algorithms
python ams_search.py

• Train RL on Custom Verification Scenarios

cd ./training
python self_improvement.py

• Open Jupyter Notebook

cd ./analyses
python3 -m notebook