SOURCE-LINKED INTELLIGENCE

Code-to-Harness: Distilling Black-Box Optimizers from Self-Play

arXiv · AI, language, vision and robotics · article · Sep 8, 2026 · UTC

Can an agent learn a numerical search strategy through executable practice and then transfer that strategy as text? We study low-budget black-box optimization, where unaided language models remain well below strong classical optimizers. During development, an agent repeatedly writes and evaluates optimizer programs. It then distills the resulting program and practice record once into a 197-word primary Harness A, which is frozen before evaluation. Harness A reduces Gemini Flash regret by 48\% in an independent $N=30$ study ($p<.001$), enters the GP-BO performance range on the practice family,

Read original source ↗ Open in workspace

recordType

paper

region

Global