Learning Evidence of Depression Symptoms via Prompt Induction

2026-04-27

AI Summary

This study addresses the challenge of depression screening under constrained clinical resources by proposing a Symptom Induction method that distills annotated samples into concise, interpretable symptom evidence guidelines. These guidelines direct large language models (LLMs) to perform sentence-level classification aligned with the 21 depressive symptoms defined in the BDI-II scale. Integrating prompt engineering, in-context learning, and rule-guided conditional classification, the approach achieves state-of-the-art weighted F1 scores across eight models from four LLM families on the BDI-Sen dataset. It substantially improves recognition consistency and generalization for low-frequency symptoms and demonstrates effective transferability of the symptom guidelines on external datasets spanning multiple disorders.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Prompt Engineering / Prompting
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• 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: Large language models for search

Abstract

Depression places substantial pressure on mental health services, and many people describe their experiences outside clinical settings in high-volume user-generated text (e.g., online forums and social media). Automatically identifying clinical symptom evidence in such text can therefore complement limited clinical capacity and scale to large populations. We address this need through sentence-level classification of 21 depression symptoms from the BDI-II questionnaire, using BDI-Sen, a dataset annotated for symptom relevance. This task is fine-grained and highly imbalanced, and we find that common LLM approaches (zero-shot, in-context learning, and fine-tuning) struggle to apply consistent relevance criteria for most symptoms. We propose Symptom Induction (SI), a novel approach which compresses labeled examples into short, interpretable guidelines that specify what counts as evidence for each symptom and uses these guidelines to condition classification. Across four LLM families and eight models, SI achieves the best overall weighted F1 on BDI-Sen, with especially large gains for infrequent symptoms. Cross-domain evaluation on an external dataset further shows that induced guidelines generalize across other diseases shared symptomatology (bipolar and eating disorders).

Problem

• depression symptoms
• user-generated text
• symptom evidence
• sentence-level classification
• mental health

Innovation

• Symptom Induction
• Depression Symptom Classification
• Prompt Engineering
• Interpretable Guidelines
• Cross-domain Generalization