BoTorch

Researchers introduced agentic Bayesian optimization, a paradigm where an LLM agent acts as the central decision maker in the BO loop while a Bayesian backend provides uncertainty-aware optimization. …