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

TL;DR

This paper introduces USP, a novel user simulator that infers implicit user profiles from interactions, enhancing realism, diversity, and consistency in dialogue simulation for training and evaluation of language models.

Contribution

We propose a framework that infers implicit user profiles from interactions and refines simulation with supervised fine-tuning and reinforcement learning, improving authenticity and diversity.

Findings

USP outperforms baselines in authenticity and diversity.

USP maintains conversation-level consistency.

Effective in evaluating LLMs on real-world benchmarks.

Abstract

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.