发表机构
University of Pennsylvania; Hark; The Chinese University of Hong Kong, Shenzhen（宾夕法尼亚大学; Hark; 香港中文大学（深圳））
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出语义范围投影进化（SSPE），将梯度投影原理从参数空间迁移至行为空间，实现语言代理共享技能的持续自我进化，在合成流和真实基准上提升跨领域能力并减轻遗忘。
AI 中文摘要
语言模型代理越来越依赖持久的自然语言技能，以超越其冻结的模型参数进行适应。然而，当共享技能从非平稳、异构的任务流中反复修订时，对新任务的改进可能会覆盖早期任务所需的程序。在持续学习中，正交梯度下降（OGD）通过将新任务的梯度投影到局部保留先前预测的子空间来解决类似的干扰问题。然而，自然语言技能的修订既没有梯度，也没有可执行此类投影的规范向量空间。我们引入了语义范围投影进化（SSPE），它将梯度投影的功能原理从参数空间转移到行为空间。SSPE将无约束的技能修订视为提议的更新，识别可能与之干扰的已获得能力，并利用观察到的增益和回归来构建兼容的修订，而不仅仅是拒绝更新。这使得一个共享技能能够在潜在和重复出现的任务上下文中进化，而无需向进化模型暴露语义领域身份。在受控的合成流和异构真实代理基准测试中，与强技能进化基线相比，SSPE提高了最终的跨领域能力并减轻了遗忘。进化后的技能在转移到不同的执行器模型后也保留了最强的平均性能。这些结果确立了语义投影作为语言代理稳定且自适应自我进化的一个有前景的原则。
英文摘要
Language-model agents increasingly rely on persistent natural-language skills to adapt beyond their frozen model parameters. When a shared skill is repeatedly revised from a non-stationary, heterogeneous task stream, however, improvements for new tasks can overwrite procedures needed for earlier ones. In continual learning, Orthogonal Gradient Descent (OGD) addresses analogous interference by projecting a new-task gradient onto a subspace that locally preserves prior predictions. Natural-language skill revisions, however, have neither gradients nor a canonical vector space in which such a projection can be performed. We introduce \emph{Semantic-Scope Projected Evolution} (SSPE), which transfers the functional principle of gradient projection from parameter space to behavior space. SSPE treats an unconstrained skill revision as a proposed update, identifies acquired capabilities with which it may interfere, and uses the observed gains and regressions to construct a compatible revision rather than merely rejecting the update. This enables one shared skill to evolve across latent and recurring task contexts without exposing semantic domain identities to the evolution model. Across controlled synthetic streams and heterogeneous real-agent benchmarks, SSPE improves final cross-domain competence and mitigates forgetting relative to strong skill-evolution baselines. The evolved skill also retains the strongest average performance after transfer to a different executor model. These results establish semantic projection as a promising principle for stable and adaptive self evolution of language agents.
