Active Learning — closing the loop

What it does

The loop every experimentalist wants: model predicts → lab measures → model updates. The Active Learning engine proposes the next batch of experiments from an existing optimization job's history, prioritizing the points where new measurements teach the model the most.

The engine

• Bayesian core — a Gaussian-process-based approach to expected information gain.
• Graceful fallback — when the full engine isn't available, a heuristic surrogate runs instead. The response always states which engine actually ran.

How to use it

Run an optimization, record a few lab results, then request the next batch. Each round of real results sharpens the next set of recommendations — the search converges on high-information experiments instead of arbitrary ones.