PRISM：Festina Lente主动性——面向风险敏感和不确定性意识的主动代理 deliberation
PRISM: Festina Lente Proactivity -- Risk-Sensitive, Uncertainty-Aware Deliberation for Proactive Agents
AI总结：
PRISM通过决策理论门控与选择性慢速推理，提升主动代理的精确性和可控性。
摘要：
主动代理必须决定不仅说什么，还要何时以及是否干预。许多现有系统依赖于脆弱的启发式方法或 indiscriminate 的长时间推理，这在控制收益与负担的权衡方面几乎没有帮助。我们将问题公式化为成本敏感的选择性干预，并提出PRISM，一种新的框架，将决策理论的门与双过程推理架构相结合。在推理过程中，代理只有在校准的用户接受概率超过由不对称成本（错过帮助和假警报）推导出的阈值时才进行干预。受“Festina Lente”（拉丁语：“慢慢快行”）的启发，我们通过接受校准和成本推导的阈值进行门控，并仅在决策边界附近调用资源密集的慢速模式，通过反事实检查进行验证，将计算集中在模糊和高风险的案例上。训练使用门控对齐、模式锁定的蒸馏：教师运行完整的PRISM流程，为未标记的交互轨迹提供密集、可执行的监督，而学生学习一个响应策略，该策略显式地与干预门解耦，以实现可调节和可审计的控制。在ProactiveBench上，PRISM将误报减少了22.78%，并在强基线之上将F1值提高了20.14%。这些结果表明，基于原理的决策理论门控，结合选择性慢速推理和对齐的蒸馏，能够产生精确、计算高效且可控的主动代理。为了促进可重复性，我们发布了代码、模型和资源，网址为https://prism-festinalente.github.io/；所有实验都使用开源的ProactiveBench基准。
英文摘要：
Proactive agents must decide not only what to say but also whether and when to intervene. Many current systems rely on brittle heuristics or indiscriminate long reasoning, which offers little control over the benefit-burden tradeoff. We formulate the problem as cost-sensitive selective intervention and present PRISM, a novel framework that couples a decision-theoretic gate with a dual-process reasoning architecture. At inference time, the agent intervenes only when a calibrated probability of user acceptance exceeds a threshold derived from asymmetric costs of missed help and false alarms. Inspired by festina lente (Latin: "make haste slowly"), we gate by an acceptance-calibrated, cost-derived threshold and invoke a resource-intensive Slow mode with counterfactual checks only near the decision boundary, concentrating computation on ambiguous and high-stakes cases. Training uses gate-aligned, schema-locked distillation: a teacher running the full PRISM pipeline provides dense, executable supervision on unlabeled interaction traces, while the student learns a response policy that is explicitly decoupled from the intervention gate to enable tunable and auditable control. On ProactiveBench, PRISM reduces false alarms by 22.78% and improves F1 by 20.14% over strong baselines. These results show that principled decision-theoretic gating, paired with selective slow reasoning and aligned distillation, yields proactive agents that are precise, computationally efficient, and controllable. To facilitate reproducibility, we release our code, models, and resources at https://prism-festinalente.github.io/; all experiments use the open-source ProactiveBench benchmark.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
