arXiv cs.AI· Lingxiao Jing, Zhengbo Jiao, Yichen Xiang, Bofan Zhu, Lingbei Meng, Yan Ni, Haorui Chen, Ran Ran, Zhirong Shen, Yuheng Yang, Chenhang Cui, Yaosong Du, Rui Huang·· 3 天前AI 评分35
RES：面向通用长时程智能体的递归环境缩放
Recursive Environment Scaling for General Long-Horizon Agents
AI 导读
研究者提出递归环境缩放（RES）框架，在保持任务目标与完成要求不变的前提下改造现有 benchmark 环境，通过 proposer 改变信息访问、资源关系与执行依赖，并由执行验证筛选候选改动、跨轮继承。
正文 · 原文
Authors:Lingxiao Jing, Zhengbo Jiao, Yichen Xiang, Bofan Zhu, Lingbei Meng, Yan Ni, Haorui Chen, Ran Ran, Zhirong Shen, Yuheng Yang, Chenhang Cui, Yaosong Du, Rui Huang
Abstract:We study environment scaling as a way to increase agent-benchmark difficulty while holding core task objectives fixed. Our framework, Recursive Environment Scaling (RES), transforms existing benchmark environments while preserving their objectives and completion requirements. A proposer changes information access, resource relations, and execution dependencies; execution-grounded checks validate the resulting candidate, and accepted changes are inherited across rounds. This produces related environments for paired evaluation and training-trajectory collection. Experiments on 294 initial tasks from three benchmarks show lower task success for all six evaluation models, with decreases of 57.14--77.14 percentage points on Claw-Eval-Live. With equal trajectory counts and identical optimization settings, fine-tuning on RES trajectories outperforms original-trajectory SFT in all nine model--benchmark comparisons by 7.00--14.61 percentage points. Ablations identify recursive inheritance and dependency composition as important contributors to increased difficulty. Together, these results support environment scaling for both agent evaluation and training.
| Comments: | 20 pages, 10 figures. Code: to be released |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.16397 [cs.AI] |
| (or arXiv:2610.16397v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.16397 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Lingxiao Jing [view email]
[v1]
Mon, 12 Oct 2026 08:39:40 UTC (1,340 KB)
来源：arXiv cs.AI · arxiv.org