Beyond Small Talk: Modeling Disclosure Trajectories in Human-AI Conversation
作者: Keith Kwok, Wanting Ning, Tej Deep Pala, Desmond C. Ong, S. Supraja, Renwen Zhang, Soujanya Poria
分类: cs.CL, cs.AI, cs.HC
发布日期: 2026-10-21
备注: 71 pages, 11 figures. Code available at https://github.com/kwokkeith/Beyond_Small_Talk
💡 一句话要点
提出TraDE框架以解决人机对话中的披露轨迹问题
🎯 匹配领域: 支柱四:生成式动作 (Generative Motion)
关键词: 人机对话 披露轨迹 情感支持 社会渗透理论 对话系统 模型评估 关系行为
📋 核心要点
- 现有方法通常只评估孤立的对话状态,无法有效分析披露历史对当前对话的影响。
- 本文提出TraDE框架,基于社会渗透理论,利用预定义的披露轨迹评估历史披露对模型行为的影响。
- 实验结果表明,模型在响应深度匹配的披露时,受先前轨迹影响,且在情感支持对话中表现出与人类支持者的显著差异。
📝 摘要(中文)
近年来,作为社会和情感支持的一种形式,对话代理的使用逐渐增加,披露过程随时间展开。然而,现有的评估方法通常只关注孤立的对话状态,难以有效分析披露历史对当前对话的影响。为此,本文提出了基于轨迹的披露评估框架(TraDE),该框架基于社会渗透理论,利用预定义的披露轨迹评估先前披露历史如何影响模型在匹配当前披露下的行为。通过TraDE,我们展示了在多个模型和关系行为基准下,基于深度匹配的披露响应会因先前轨迹的不同而有所差异,尤其是在模型接受或挑战用户框架时。这些差异与对深度披露的累积接触有关,而非单一脆弱时刻或披露顺序。
🔬 方法详解
问题定义:本文旨在解决当前人机对话中披露历史与当前披露状态相混淆的问题,现有方法无法有效捕捉披露轨迹对对话行为的影响。
核心思路:提出TraDE框架,通过预定义的披露轨迹来评估历史披露如何影响模型在当前披露下的反应,强调披露的时间维度。
技术框架:TraDE框架包括数据收集、披露轨迹定义、模型训练与评估等主要模块,利用社会渗透理论指导模型行为分析。
关键创新:TraDE的核心创新在于系统性地揭示历史依赖的关系行为,超越了传统孤立状态评估的局限,提供了更全面的对话理解。
关键设计:在模型训练中,采用了特定的损失函数来优化披露深度匹配,同时设计了多层次的网络结构以捕捉披露轨迹的复杂性。通过这些设计,模型能够更好地反映人类的情感支持行为。
🖼️ 关键图片
📊 实验亮点
实验结果显示,模型在情感支持对话中,响应的差异性与先前披露轨迹密切相关,尤其是在用户减少披露时,模型的表现与人类支持者存在显著差异。这一发现为未来的对话系统设计提供了重要的参考。
🎯 应用场景
该研究的潜在应用领域包括情感支持系统、智能客服以及社交机器人等。通过更好地理解人机对话中的披露轨迹,能够提升对话代理的响应质量和用户体验,进而在心理健康支持和社交互动中发挥重要作用。
📄 摘要(原文)
Recently, we see a rise in conversational agents as a form of social and emotional support, where disclosure will unfold over time. Yet, current evaluations of these agents typically examine isolated conversational states. In analyses of naturally occurring conversations, disclosure history and current disclosure are confounded which tends to make their effects difficult to isolate. We thus introduce Trajectory-based Disclosure Evaluation (TraDE), an evaluation framework grounded in Social Penetration Theory that utilizes predefined disclosure trajectories and evaluates how preceding disclosure history shapes model behavior under matched current disclosure. Through the use of TraDE, we demonstrate across multiple models and relational behavior benchmarks that responses to disclosures matched on depth can differ depending on the preceding trajectory, particularly in whether models accept or challenge the user's framing. We also observed that these differences track accumulated exposure to deeper disclosure rather than any single moment of vulnerability or the order of disclosures. It also tends to persists for several turns after disclosure returns to shallow levels. For several behaviors, we also demonstrate that models diverge most consistently from human supporters in the Emotional Support Conversation corpus (ESConv) not when users open up but rather when they step back in their disclosure. TraDE thus provides a systematic way to expose history-dependent relational behavior that evaluations of isolated conversational states would otherwise miss.