Bayesian Optimizer

Bayesian Optimization (Self-Driving Lab)

The Bayesian Optimizer allows agents to efficiently explore a parameter space to maximize a target metric (yield, purity, binding affinity) with minimal experiments. It uses Gaussian Processes to model uncertainty and the Upper Confidence Bound (UCB) acquisition function.

When to Use This Skill

• When experiments are expensive or time-consuming.
• To autonomously tune hyperparameters for a machine learning model.
• To optimize reaction conditions (temperature, pH, concentration).

Core Capabilities

1. Next Step Proposal: Suggests the next best experiment parameters.
2. Surrogate Modeling: Predicts outcomes for untested parameters.
3. Exploration/Exploitation: Balances trying new things vs. refining known good results.

Workflow

1. Input: History of past experiments (params -> results) and bounds.
2. Process: Fits a Gaussian Process to the data.
3. Output: Returns the parameters for the next experiment.

Example Usage

User: "Given these past results, what temperature and pH should I try next?"

Agent Action:

python3 Skills/Mathematics/Probability_Statistics/bayesian_optimization.py \
    --history "[[20, 7.0, 0.5], [25, 6.5, 0.6]]" \
    --bounds "[[10, 40], [5, 9]]" \
    --output next_experiment.json