先识你，再懂你：通过隐式画像构建类人用户模拟器

摘要

用户模拟器对于复现人类与对话系统的交互至关重要，它既支持协作训练，也支持自动评估，尤其是在大型语言模型（LLMs）的应用中。然而，现有的模拟器往往仅依赖于文本话语，忽视了诸如个性、说话风格和目标等隐含的用户特质。相比之下，基于人物角色的方法因依赖预定义的名人或原型档案而缺乏普适性。为应对这些挑战，我们提出了带有隐含用户档案的用户模拟器（USP），该框架能够从人机对话中推断出隐含的用户档案，并利用这些档案生成更加个性化和真实的对话。我们首先开发了一个基于LLM的提取器，配备了一套全面的档案模式。随后，通过条件监督微调和循环一致性的强化学习，我们在话语和对话两个层面上对模拟器进行了优化。最后，我们采用多样化的档案采样器来捕捉现实世界用户档案的分布。实验结果表明，USP在真实性和多样性方面均优于强基线，同时在一致性方面也达到了可比的表现。此外，基于USP的动态多轮评估与主流基准高度一致，证明了其在现实应用中的有效性。

English

User simulators are crucial for replicating human interactions with dialogue systems, supporting both collaborative training and automatic evaluation, especially for large language models (LLMs). However, existing simulators often rely solely on text utterances, missing implicit user traits such as personality, speaking style, and goals. In contrast, persona-based methods lack generalizability, as they depend on predefined profiles of famous individuals or archetypes. To address these challenges, we propose User Simulator with implicit Profiles (USP), a framework that infers implicit user profiles from human-machine conversations and uses them to generate more personalized and realistic dialogues. We first develop an LLM-driven extractor with a comprehensive profile schema. Then, we refine the simulation through conditional supervised fine-tuning and reinforcement learning with cycle consistency, optimizing it at both the utterance and conversation levels. Finally, we adopt a diverse profile sampler to capture the distribution of real-world user profiles. Experimental results demonstrate that USP outperforms strong baselines in terms of authenticity and diversity while achieving comparable performance in consistency. Furthermore, dynamic multi-turn evaluations based on USP strongly align with mainstream benchmarks, demonstrating its effectiveness in real-world applications.

先识你，再懂你：通过隐式画像构建类人用户模拟器

Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles

摘要

Summary

Support