Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912455343472640 |
|---|---|
| author | Wang, Kuang Li, Xianfei Yang, Shenghao Zhou, Li Jiang, Feng Li, Haizhou |
| author_facet | Wang, Kuang Li, Xianfei Yang, Shenghao Zhou, Li Jiang, Feng Li, Haizhou |
| contents | 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, current role-playing methods face challenges such as a lack of utterance-level authenticity and user-level diversity, often hindered by role confusion and dependence on predefined profiles of well-known figures. In contrast, direct simulation focuses solely on text, neglecting implicit user traits like personality and conversation-level consistency. To address these issues, we introduce the User Simulator with Implicit Profiles (USP), a framework that infers implicit user profiles from human-machine interactions to simulate personalized and realistic dialogues. We first develop an LLM-driven extractor with a comprehensive profile schema, then refine the simulation using conditional supervised fine-tuning and reinforcement learning with cycle consistency, optimizing at both the utterance and conversation levels. Finally, a diverse profile sampler captures the distribution of real-world user profiles. Experimental results show that USP outperforms strong baselines in terms of authenticity and diversity while maintaining comparable consistency. Additionally, using USP to evaluate LLM on dynamic multi-turn aligns well with mainstream benchmarks, demonstrating its effectiveness in real-world applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_18968 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles Wang, Kuang Li, Xianfei Yang, Shenghao Zhou, Li Jiang, Feng Li, Haizhou Computation and Language 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, current role-playing methods face challenges such as a lack of utterance-level authenticity and user-level diversity, often hindered by role confusion and dependence on predefined profiles of well-known figures. In contrast, direct simulation focuses solely on text, neglecting implicit user traits like personality and conversation-level consistency. To address these issues, we introduce the User Simulator with Implicit Profiles (USP), a framework that infers implicit user profiles from human-machine interactions to simulate personalized and realistic dialogues. We first develop an LLM-driven extractor with a comprehensive profile schema, then refine the simulation using conditional supervised fine-tuning and reinforcement learning with cycle consistency, optimizing at both the utterance and conversation levels. Finally, a diverse profile sampler captures the distribution of real-world user profiles. Experimental results show that USP outperforms strong baselines in terms of authenticity and diversity while maintaining comparable consistency. Additionally, using USP to evaluate LLM on dynamic multi-turn aligns well with mainstream benchmarks, demonstrating its effectiveness in real-world applications. |
| title | Know You First and Be You Better: Modeling Human-Like User Simulators via Implicit Profiles |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2502.18968 |