阿里发布 G-CAST:让电商智能体持续自我训练的治理框架
G-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving Demands
阿里把自家 1688 B2B 采购智能体的持续训练方法写成了论文,核心是怎么让线上智能体天天拿生产数据自动迭代,还不忘旧功能。做 LLM 智能体运维、线上模型更新的工程团队可以参考它的检查点切换和奖励路由设计。
阿里巴巴在 arXiv 发布论文《G-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving Demands》,提出一套面向电商智能体的持续自我训练系统。该系统已被部署在 1688 平台的 B2B 采购智能体 Yuancai(cai.1688.com)上。
论文针对两个核心痛点:一是需求漂移(demand drift)会让一次性训练的模型跟不上新品类和新采购场景的请求分布变化;二是执行框架(harness)中工具、技能和操作流程不断更新,模型既要学会使用新能力又不能丢掉用户依赖的旧行为。G-CAST 将每日生产数据在需求和能力的治理契约下转化为 reward-routed、entropy-prioritized 的训练数据,并通过 governed reward evolution 演化奖励函数。
模型更新由基于累积评估的受控检查点切换(governed checkpoint transition)把关。在真实生产电商智能体上的实验显示,G-CAST 相比强自主 post-training 基线,对新需求的适应更强,同时更好地保留了既有行为。消融实验证明数据编排与奖励演化两个机制均对效果有实质贡献。
G-CAST: Governed Continual Self-Training for E-Commerce Agents under Evolving Demands
With the rapid advancement of Large Language Models (LLMs), an increasing number of users are turning to LLM-powered agents for conversational shopping in e-commerce. However, the user needs continue to change as new product categories, procurement scenarios, and feature requests emerge. At the same time, the product's execution harness-including its tools, skills, and operating procedures-continues to evolve to support increasingly sophisticated interactions. These changes create two practical challenges: 1) Demand drift alters the distribution, intent, and complexity of user requests, causing a model trained once to fall behind; 2) Harness evolution requires the model to use newly introduced capabilities correctly without losing behavior that users already rely on. Periodic manual retraining is too delayed and labor-intensive for this setting. To address these challenges, we present G-CAST, a governed continual self-training system that converts daily production evidence under evolving demand and capability contracts into reward-routed, entropy-prioritized training data and governed reward evolution. Governed checkpoint transition under cumulative evaluation controls each model update, enabling the agent to learn newly introduced behaviors without sacrificing previously accepted capabilities. Experiments on a real production e-commerce agent show that G-CAST achieves stronger adaptation to newly introduced demand while better preserving previously accepted behavior compared with a strong autonomous post-training baseline. Ablations further show that both reward-routed, entropy-prioritized data orchestration and governed reward evolution are important to these gains. G-CAST is now deployed in Yuancai(http://cai.1688.com/), a B2B procurement agent on Alibaba's 1688 platform.
本条情报归属 AITOP「论文」核心栏目,重点聚焦中国 AI 产业与国产大模型生态。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」「中美大模型技术代差与商业化对照」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:G-CAST、Yuancai、1688、智能体,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「arXiv cs.AI」(@Yuechun Yu, Wenzheng Kong, Xiaofeng Mao, Jin Zhang),经 AITOP 多维评分管线结构化提炼。