Salvato in:
Dettagli Bibliografici
Autori principali: Li, Hanchen, He, Runyuan, Zhang, Qizheng, Ji, Changxiu, Mang, Qiuyang, Chen, Xiaokun, Agrawal, Lakshya A, Liao, Wei-Liang, Yang, Eric, Cheung, Alvin, Zou, James, Olukotun, Kunle, Stoica, Ion, Gonzalez, Joseph E.
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2604.04247
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918429838016512
author Li, Hanchen
He, Runyuan
Zhang, Qizheng
Ji, Changxiu
Mang, Qiuyang
Chen, Xiaokun
Agrawal, Lakshya A
Liao, Wei-Liang
Yang, Eric
Cheung, Alvin
Zou, James
Olukotun, Kunle
Stoica, Ion
Gonzalez, Joseph E.
author_facet Li, Hanchen
He, Runyuan
Zhang, Qizheng
Ji, Changxiu
Mang, Qiuyang
Chen, Xiaokun
Agrawal, Lakshya A
Liao, Wei-Liang
Yang, Eric
Cheung, Alvin
Zou, James
Olukotun, Kunle
Stoica, Ion
Gonzalez, Joseph E.
contents Recent advances in prompt learning allow large language model agents to acquire task-relevant knowledge from inference-time context without parameter changes. For example, existing methods (like ACE or GEPA) can learn system prompts to improve accuracy based on previous agent runs. However, these methods primarily focus on single-agent or low-parallelism settings. This fundamentally limits their ability to efficiently learn from a large set of collected agentic traces. It would be efficient and beneficial to run prompt learning in parallel to accommodate the growing trend of learning from many agentic traces or parallel agent executions. Yet without a principled strategy for scaling, current methods suffer from quality degradation with high parallelism. To improve both the efficiency and quality of prompt learning, we propose Combee, a novel framework to scale parallel prompt learning for self-improving agents. Combee speeds up learning and enables running many agents in parallel while learning from their aggregate traces without quality degradation. To achieve this, Combee leverages parallel scans and employs an augmented shuffle mechanism; Combee also introduces a dynamic batch size controller to balance quality and delay. Evaluations on AppWorld, Terminal-Bench, Formula, and FiNER demonstrate that Combee achieves up to 17x speedup over previous methods with comparable or better accuracy and equivalent cost.
format Preprint
id arxiv_https___arxiv_org_abs_2604_04247
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
Li, Hanchen
He, Runyuan
Zhang, Qizheng
Ji, Changxiu
Mang, Qiuyang
Chen, Xiaokun
Agrawal, Lakshya A
Liao, Wei-Liang
Yang, Eric
Cheung, Alvin
Zou, James
Olukotun, Kunle
Stoica, Ion
Gonzalez, Joseph E.
Artificial Intelligence
Computation and Language
Machine Learning
Recent advances in prompt learning allow large language model agents to acquire task-relevant knowledge from inference-time context without parameter changes. For example, existing methods (like ACE or GEPA) can learn system prompts to improve accuracy based on previous agent runs. However, these methods primarily focus on single-agent or low-parallelism settings. This fundamentally limits their ability to efficiently learn from a large set of collected agentic traces. It would be efficient and beneficial to run prompt learning in parallel to accommodate the growing trend of learning from many agentic traces or parallel agent executions. Yet without a principled strategy for scaling, current methods suffer from quality degradation with high parallelism. To improve both the efficiency and quality of prompt learning, we propose Combee, a novel framework to scale parallel prompt learning for self-improving agents. Combee speeds up learning and enables running many agents in parallel while learning from their aggregate traces without quality degradation. To achieve this, Combee leverages parallel scans and employs an augmented shuffle mechanism; Combee also introduces a dynamic batch size controller to balance quality and delay. Evaluations on AppWorld, Terminal-Bench, Formula, and FiNER demonstrate that Combee achieves up to 17x speedup over previous methods with comparable or better accuracy and equivalent cost.
title Combee: Scaling Prompt Learning for Self-Improving Language Model Agents
topic Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2604.04247