research
The Search Tree Had to Earn Every New Branch
ExTS treats expansion as a value-of-information decision when evaluations and model calls are scarce.
Summary
ExTS treats expansion as a value-of-information decision when evaluations and model calls are scarce.
The policy combines sharper reward separation, a virtual child that estimates the value of branching and quality-conditioned expansion. Across prompt optimization, code generation, molecular structure elucidation and workflow optimization, one fixed configuration produced a reported average relative gain of 5.5 percent over task-specific tree-search baselines.
Why it matters
ExTS treats expansion as a value-of-information decision when evaluations and model calls are scarce.
Limits and context
No additional limitation was separately recorded.
Key claims
ExTS treats expansion as a value-of-information decision when evaluations and model calls are scarce.
Evidence: source-2026-08-26-013
Sources
- arXiv preprint 2608.23848arXiv · primary research
Corrections
No corrections have been recorded for this story.