PulseAugur
EN
LIVE 06:07:24

Language models learn search strategies via practice, transferable as text

By PulseAugur Editorial · [1 sources] ·

Researchers have developed a method called Code-to-Harness that enables language models to learn numerical search strategies through practice and then distill these strategies into text. This approach significantly reduces regret in black-box optimization tasks, outperforming unaided language models and rivaling classical optimizers. The distilled text-based strategies are transferable across different models, including Gemini Flash and Claude Sonnet, and have shown effectiveness on real-world production benchmarks. AI

IMPACT This method could enable more efficient and adaptable optimization strategies for AI agents across various applications.

RANK_REASON The cluster contains an academic paper detailing a novel method for training language models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Language models learn search strategies via practice, transferable as text

How we ranked this

Signal score
35 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
The cluster contains an academic paper detailing a novel method for training language models. [lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. arXiv cs.LG TIER_1 English(EN) · Yi Wu, Zheng Ren, Zhiyu Hu, Haochen Wang, Daryl Chang, Li Wei, Ting Wang, Zhen Li, Pooja Gupta, Nitin Jindal, Lukasz Heldt ·

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

    arXiv:2609.09468v2 Announce Type: replace Abstract: 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 opt…

RELATED ENTITIES

RELATED TOPICS