BEDA：信念估计作为执行策略性对话行为的概率约束

摘要

战略对话要求智能体执行不同的对话行为，信念估计对此至关重要。虽然现有研究能实现较准确的信念估计，但缺乏在生成过程中运用这些信念的规范机制。我们通过以下方式填补这一空白：首先将对抗性和协调性这两类核心行为形式化，进而通过对智能体生成内容施加概率约束来实现操作化。基于此思路，我们构建了BEDA框架，该框架包含世界状态集合、负责信念估计的信念估计器，以及根据推断信念选择行为并生成一致性话语的条件生成器。在条件型守护者-窃贼（CKBG，对抗性）、共同好友（MF，合作性）和卡西诺（协商性）三类场景中，BEDA均持续超越强基线模型：在CKBG任务中，不同骨干模型上的成功率提升至少5.0个百分点，使用GPT-4.1-nano时提升达20.6个百分点；在共同好友任务中平均提升9.3个百分点；在卡西诺任务中达成了相对于所有基线模型的最优协商方案。这些结果表明，将信念估计转化为约束条件，为可靠的战略对话提供了一种简洁通用的实现机制。

English

Strategic dialogue requires agents to execute distinct dialogue acts, for which belief estimation is essential. While prior work often estimates beliefs accurately, it lacks a principled mechanism to use those beliefs during generation. We bridge this gap by first formalizing two core acts Adversarial and Alignment, and by operationalizing them via probabilistic constraints on what an agent may generate. We instantiate this idea in BEDA, a framework that consists of the world set, the belief estimator for belief estimation, and the conditional generator that selects acts and realizes utterances consistent with the inferred beliefs. Across three settings, Conditional Keeper Burglar (CKBG, adversarial), Mutual Friends (MF, cooperative), and CaSiNo (negotiation), BEDA consistently outperforms strong baselines: on CKBG it improves success rate by at least 5.0 points across backbones and by 20.6 points with GPT-4.1-nano; on Mutual Friends it achieves an average improvement of 9.3 points; and on CaSiNo it achieves the optimal deal relative to all baselines. These results indicate that casting belief estimation as constraints provides a simple, general mechanism for reliable strategic dialogue.

BEDA：信念估计作为执行策略性对话行为的概率约束

BEDA: Belief Estimation as Probabilistic Constraints for Performing Strategic Dialogue Acts

摘要

Support