Language models improve by simulating users’ mental states
Mind2Dialogue: Training Human-Aware Language Models by Simulating User Mental States
Computation and LanguageMachine Learning
Summary
Helping computers understand people better can make conversations and collaborations more useful. But it's hard to train computers on what people really think or want because we can't see those thoughts directly. The authors created a system that pretends to be a user with changing thoughts and goals to teach language models how to respond better. They showed that models trained this way follow user preferences much more and can reason about beliefs and actions more accurately.
What this means in practice
- •For customer service teams: Develop chatbots that better anticipate and respond to customers’ unspoken needs and preferences in long conversations.$Commercial implications: Improves chatbot quality for businesses by enabling assistants to understand and adapt to user goals, enhancing customer satisfaction.
- •For personal assistant developers: Create virtual assistants that adapt responses based on inferred user intentions and evolving preferences without explicit user input.$Commercial implications: Enables more natural and helpful virtual assistants that can better support users over time by simulating mental state understanding.
Authors
Zixuan Wang, Yufan Zhou, Jinzhou Tang, Xinle Yu, Chengjun Wu, Lyumanshan Ye, Zhaoxiang Feng, Letian Peng, Adyasha Patra, Fan Bai, Enze Ma, Zhengding Hu, Jianyang Gu, Zhao Wang, Yufei Ding, Jingbo Shang, Tianmin Shu, Zhiting Hu, Zhen Wang
Abstract
As language models become more capable, long-term collaboration in learning, reasoning, and decision-making calls for a deeper understanding of the people they serve. Yet training such human-aware language models faces a fundamental supervision gap because current datasets for LLM assistant training contain few if any well-informed responses explicitly grounded in users' unspoken beliefs and goals. Scaling such supervision is inherently constrained, as users' underlying states are not directly observable. We thus propose the Mind2Dialogue framework to mitigate this gap by simulating users' mental states and turning them into privileged supervision for human-aware training. Specifically, we first propose a psychology-guided simulator that preserves personal characteristics while updating mental states through interaction to generate coherent conversations. The key idea is to enforce a shared evolving mental state that drives user behavior and guides an Oracle assistant's responses. Our privileged distillation then trains models on the Oracle's well-informed responses to assist users without direct access to their mental states at deployment. Moreover, we propose to evaluate human-aware learning by combining personalization and theory of mind, examining how models understand people and act on that understanding. Training on the full Mind2Dialogue corpus improves every reported personalization metric over the corresponding Qwen, Llama, and OLMo instruction-tuned baselines, including gains of 26.6 to 40.9 percentage points in preference-following generation. The gains extend to belief and action reasoning on Qwen and Llama, beyond personalized assistance. Looking forward, Mind2Dialogue makes user simulation a foundation for genuine AI collaborators that understand beliefs and intentions behind people's words and support their long-term goals across education, work, and everyday life.