arXiv cs.AI· Renhao Li, Xinjie Shen, Yang Su, Jianhong Tu, Derek F. Wong, Dayiheng Liu, Min Yang·· 2 天前AI 评分38
MetaExp：通过可验证跨任务环境扩展让模型从经验中自我改进
MetaExp: Learning to Self-Improve through Experience via Verifiable Cross-Task Environment Scaling
AI 导读
MetaExp 通过可验证任务族训练模型的经验学习能力，在 Qwen3.5-122B-A10B 上用 1,000 个任务族的 4,956 条轨迹微调，下游参数冻结下 SpreadsheetBench 提升 +8.5 分、AutomationBench 提升 +14.8 分，APEX-Agents 峰值 +2.3。
正文 · 原文
Abstract:Language-model agents can accumulate interaction feedback, but reliably turning that feedback into reusable experience remains challenging. Existing methods either optimize external experience scaffolds or train experience-related behavior on specific task distributions, leaving open whether experience learning itself can become a transferable model capability. Training such a capability calls for supervision beyond isolated tasks, which mainly support in-episode problem solving rather than revealing what knowledge remains useful across episodes. In this work, we introduce MetaExp, a scalable framework that trains this capability through verifiable task families with controlled cross-task structure. Each family preserves a reusable factor while varying task-specific realizations, enabling transfer without direct solution reuse. We propose two complementary designs: Shared-Tool families keep black-box tools fixed across tasks to elicit operational experience, while Shared-Workflow families preserve latent procedures across changing realizations to elicit procedural experience. We scale environments under both designs, collect and filter sequential interaction trajectories, and fine-tune Qwen3.5-122B-A10B on 4,956 trajectories from 1,000 task families. With parameters frozen during downstream evolution, MetaExp improves by +8.5 points on SpreadsheetBench, peaks at +2.3 on APEX-Agents, and improves by +14.8 on AutomationBench. The gains persist under domain-excluded training, supporting experience learning as a transferable capability for continued non-parametric improvement after deployment.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.21492 [cs.AI] |
| (or arXiv:2610.21492v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.21492 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Renhao Li [view email]
[v1]
Thu, 15 Oct 2026 08:29:01 UTC (492 KB)
来源：arXiv cs.AI · arxiv.org