Aspire：模型能否从模糊目标中自我进化？

Aspire: Can Models Self-Evolve from Vague Goals?

August 31, 2026

作者: Yuhao Wu, Jingyuan Zhang, Jiajun Shi, Yuxuan Zhang, Xinping Lei, Junting Zhou, Zexuan Wang, Yuchen Wu, Huan Zhou, Duo Wang, Yinzhu Piao, Yongchang Peng, Yunfeng Shi, Jin Chen, Zuo Wang, Jinkai Liu, Jiaheng Liu, Wenxuan Zhang, Shen Yan, Wenhao Huang, Ge Zhang

cs.AI

摘要

人类许多重要的学习形式始于模糊的目标，例如“成为更好的物理学家”或“在研究方面取得进步”。学习者必须解读目标、识别能力差距、决定如何学习，并判断自己是否真正有所提升。相比之下，现有的关于大语言模型（LLM）自我进化的研究通常始于由人类明确设定的任务和评估指标，将自我进化简化为优化一个显式目标，而非自主决定学什么和怎么学。我们提出了ASPERE——一个面向模糊目标驱动型自我进化的基准测试。ASPERE仅提供一个自然语言形式的能力目标，而下游评估任务保持隐藏。智能体必须通过选择数据和更新方法、构建训练与验证信号、以及决定何时进行评估来将目标具体化。ASPERE在统一的交互式环境中同时支持模型权重进化和智能体框架（harness）进化，并在由专家撰写的、涵盖六个目标的520个隐藏条目上评估最终系统。我们的实验表明，模糊目标将搜索努力转向目标解读。当前的智能体能够常规性地完成训练和框架编辑循环，但权重层面的提升仍然稀少且不稳定，且进化得到的最强框架仍低于工程化设计的Qwen-Agent参考实现。智能体经常在数据不匹配的情况下进行训练，并过于相信狭隘的自我评估，导致局部提升无法迁移到隐藏评估中，而持续的搜索和训练甚至可能抹去先前取得的改进。

English

Many important forms of human learning begin with a vague goal, such as "become a better physicist" or "improve at research." Learners must interpret the goal, identify capability gaps, decide how to learn, and determine whether they have actually improved. In contrast, existing work on LLM self-evolution typically begins with tasks and evaluation metrics specified by humans, reducing self-evolution to optimizing an explicit objective rather than deciding what and how to learn. We introduce ASPIRE, a benchmark for vague-goal-driven self-evolution. ASPIRE provides only a natural-language capability goal while downstream evaluation tasks remain hidden. The agent must operationalize the goal by choosing data and update methods, constructing training and validation signals, and deciding when to evaluate. ASPIRE supports both model-weight and agent-harness evolution in a unified interactive environment and evaluates the resulting systems on a hidden, expert-authored set of 520 items spanning six goals. Our experiments show that vague goals redirect search effort toward goal interpretation. Current agents routinely complete training and harness-editing loops, but weight-level gains remain sparse and unstable, and the strongest evolved harness remains below the engineered Qwen-Agent reference. Agents often train on mismatched data and trust narrow self-evaluations, so local gains fail to transfer to hidden evaluation and continued search and training can erase earlier improvements.