Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Su, Hongyuan, Zheng, Yu, Li, Yong
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2602.02597
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917243773779968
author Su, Hongyuan
Zheng, Yu
Li, Yong
author_facet Su, Hongyuan
Zheng, Yu
Li, Yong
contents Large language models are transforming systems research by automating the discovery of performance-critical algorithms for computer systems. Despite plausible codes generated by LLMs, producing solutions that meet the stringent correctness and performance requirements of systems demands iterative optimization. Test-time reinforcement learning offers high search efficiency but requires parameter updates infeasible under API-only access, while existing training-free evolutionary methods suffer from inefficient context utilization and undirected search. We introduce ContextEvolve, a multi-agent framework that achieves RL-level search efficiency under strict parameter-blind constraints by decomposing optimization context into three orthogonal dimensions: a Summarizer Agent condenses semantic state via code-to-language abstraction, a Navigator Agent distills optimization direction from trajectory analysis, and a Sampler Agent curates experience distribution through prioritized exemplar retrieval. This orchestration forms a functional isomorphism with RL-mapping to state representation, policy gradient, and experience replay-enabling principled optimization in a textual latent space. On the ADRS benchmark, ContextEvolve outperforms state-of-the-art baselines by 33.3% while reducing token consumption by 29.0%. Codes for our work are released at https://anonymous.4open.science/r/ContextEvolve-ACC
format Preprint
id arxiv_https___arxiv_org_abs_2602_02597
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ContextEvolve: Multi-Agent Context Compression for Systems Code Optimization
Su, Hongyuan
Zheng, Yu
Li, Yong
Machine Learning
Artificial Intelligence
Large language models are transforming systems research by automating the discovery of performance-critical algorithms for computer systems. Despite plausible codes generated by LLMs, producing solutions that meet the stringent correctness and performance requirements of systems demands iterative optimization. Test-time reinforcement learning offers high search efficiency but requires parameter updates infeasible under API-only access, while existing training-free evolutionary methods suffer from inefficient context utilization and undirected search. We introduce ContextEvolve, a multi-agent framework that achieves RL-level search efficiency under strict parameter-blind constraints by decomposing optimization context into three orthogonal dimensions: a Summarizer Agent condenses semantic state via code-to-language abstraction, a Navigator Agent distills optimization direction from trajectory analysis, and a Sampler Agent curates experience distribution through prioritized exemplar retrieval. This orchestration forms a functional isomorphism with RL-mapping to state representation, policy gradient, and experience replay-enabling principled optimization in a textual latent space. On the ADRS benchmark, ContextEvolve outperforms state-of-the-art baselines by 33.3% while reducing token consumption by 29.0%. Codes for our work are released at https://anonymous.4open.science/r/ContextEvolve-ACC
title ContextEvolve: Multi-Agent Context Compression for Systems Code Optimization
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2602.02597