NASA’s new dark energy space telescope can also detect killer asteroids
technologyreview.com·3h ago
AI Research
arXiv CS.AI·Jun 10·5 min read
Dep-LLM: Training-Free Depression Diagnosis via Evidence-Guided Structured Multi-factor with Reliable LLM Reasoning
TL;DR
Automatic Depression Detection (ADD) faces challenges due to complex clues in lengthy clinical interviews and limited labeled data. Dep-LLM is a training-free framework that uses off-the-shelf large language models (LLMs) to analyze dialogues and produce reliable diagnoses.
✦ Why It Matters
Engineers can implement Dep-LLM for efficient and reliable depression diagnosis without extensive training data.
Key Takeaways
- Explore training-free frameworks for mental health applications.
- Utilize Chain-of-Thought reasoning for complex data analysis.
- Implement confidence modulation techniques to enhance model reliability.
- Evaluate performance against existing models using standard metrics.
Full Summary
Automatic Depression Detection (ADD) is crucial for mental health but is hindered by the complexity of interpreting long clinical interviews and the lack of labeled data due to privacy concerns. Dep-LLM is introduced as a training-free framework that leverages existing large language models (LLMs) to mimic the reasoning process of psychiatrists.
It operates in three stages: first, it uses a Chain-of-Thought (CoT) module to break down dialogues into five clinically relevant themes, generating evidence-based rationales. Next, a Confidence Analysis and Modulation module assesses the reliability of these rationales, enhancing trustworthy signals while minimizing uncertainty.
Finally, a Collaborative Multi-factor Prediction module integrates these signals to produce a final diagnosis. Experiments on the DAIC-WOZ and E-DAIC datasets show that Dep-LLM surpasses zero-shot baselines across 21 LLMs and outperforms state-of-the-art supervised models, demonstrating its effectiveness and generalizability.