Closed-loop molecular discovery agent combining Bayesian optimization (BoTorch), a Gymnasium environment, and a simulated QED + synthetic-accessibility oracle for autonomous drug-candidate selection.
python reinforcement-learning cheminformatics drug-discovery rdkit gaussian-processes bayesian-optimization active-learning gymnasium mlflow botorch molecular-discovery closed-loop-optimization
-
Updated
Jul 4, 2026