发表机构
UniCone Team（UniCone团队）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究聚焦主动服务智能体，构建统一决策框架，梳理现有方法，归一化决策单元与证据描述符，形式化评估指标，指出离线分类性能及长期记忆并非主动服务的关键，可靠主动服务需满足特定条件。
AI 中文摘要
大型语言模型智能体可进行规划、调用工具及修改外部状态，但多数系统仍将明确的用户指令作为固定起点。主动服务将决策环节前置：智能体需从不完整的环境和用户信号中推断服务机会，在保持沉默、询问、协助、弃权（不执行）及采取行动间做出选择，并需考虑中断、误解、过度干预及隐私成本。本综述给出以主动性为核心的可操作定义，将该问题形式化为受授权和风险约束的部分可观测序贯决策过程，该形式化在结构化动作中表示时机、内容与交付方式，同时明确等待的期权价值、询问的决策价值及反馈引发的状态变化。在此基础上，我们沿一条决策流程（状态与需求估计、干预门控、动作构建、反馈适配）梳理现有方法，并将规定式、预测式、基于模型及回报优化机制描述为非排他性的策略构建组件。我们还在流式对话、屏幕、视频、软件工程及人机协作资源中对决策单元和三轴证据描述符进行归一化，并形式化触发、时机、校准、用户负担、安全及策略价值的指标。综合分析表明，仅离线分类性能无法预测部署收益，长期记忆并非主动性的定义条件，可靠的主动服务反而需要校准的增量干预价值、可验证的授权、可恢复的执行及反事实证据。
英文摘要
Large language model agents can plan, invoke tools, and modify external states, yet most systems still take an explicit user instruction as a fixed starting point. Proactive service moves the decision upstream: an agent must infer service opportunities from incomplete environmental and user signals, choose among remaining silent, asking, assisting, and acting, and account for interruption, misunderstanding, overreach, and privacy costs. This survey gives an operational definition centered on initiative and formulates the problem as a partially observable sequential decision process constrained by authorization and risk. The formulation represents timing, content, and delivery within one structured action, while making explicit the option value of waiting, the decision value of questions, and feedback-induced state changes. On this basis, we organize existing methods along one decision pipeline (state and need estimation, intervention gating, action construction, and feedback adaptation) and describe prescribed, predictive, model based, and return optimizing mechanisms as nonexclusive policy-construction components. We further normalize decision units and three-axis evidence descriptors across streaming dialogue, screen, video, software-engineering, and human-agent collaboration resources, and formalize metrics for triggering, timing, calibration, user burden, safety, and policy value. The synthesis shows why offline classification performance alone does not predict deployment benefit and why long-term memory is not a defining condition of proactivity. Reliable proactive service instead requires calibrated incremental intervention value, verifiable authorization, recoverable execution, and counterfactual evidence.
