Enhancing Depression-Diagnosis-Oriented Chat with Psychological State Tracking

2024-03-12

arXiv.org

AI Summary

Existing depression diagnosis dialogue systems struggle to dynamically track patients’ evolving psychological states and lack structured guidance frameworks, resulting in low clinical efficiency and insufficient realism. To address this, we propose an Explicit Psychological State Tracking (POST) mechanism that, for the first time, integrates a structured theory-of-mind model into large language model (LLM)-based dialogue systems. POST establishes a four-dimensional dynamic representation—comprising stage, information, summary, and next-intent—to enable interpretable, intervention-aware, diagnosis-oriented dialogue guidance. Our method jointly optimizes psychological state sequence modeling, multi-stage policy control, and response generation. On benchmark datasets, POST achieves state-of-the-art performance across key subtasks—including symptom identification, diagnostic stage progression, and response relevance—yielding a 12.6% improvement in diagnostic recommendation accuracy.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Intelligent Robots: State Estimation

Application Category

• User Modeling, Personalization and Recommendation: Psychology-informed user models and recommender systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Depression-diagnosis-oriented chat aims to guide patients in self-expression to collect key symptoms for depression detection. Recent work focuses on combining task-oriented dialogue and chitchat to simulate the interview-based depression diagnosis. Whereas, these methods can not well capture the changing information, feelings, or symptoms of the patient during dialogues. Moreover, no explicit framework has been explored to guide the dialogue, which results in some useless communications that affect the experience. In this paper, we propose to integrate Psychological State Tracking (POST) within the large language model (LLM) to explicitly guide depression-diagnosis-oriented chat. Specifically, the state is adapted from a psychological theoretical model, which consists of four components, namely Stage, Information, Summary and Next. We fine-tune an LLM model to generate the dynamic psychological state, which is further used to assist response generation at each turn to simulate the psychiatrist. Experimental results on the existing benchmark show that our proposed method boosts the performance of all subtasks in depression-diagnosis-oriented chat.

Problem

• Tracking dynamic psychological states during diagnostic dialogues
• Improving depression diagnosis through guided conversational frameworks
• Reducing irrelevant communications in mental health assessment chats

Innovation

• Integrates Psychological State Tracking with LLM
• Fine-tunes LLM for dynamic state generation
• Uses state components to guide dialogue