Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Multi-aspect Depression Severity Assessment via Inductive Dialogue System

Chaebin Lee, Seungyeon Seo, Heejin Do, Gary Geunbae Lee

TL;DR

A novel task of multi-aspect depression severity assessment via an inductive dialogue system (MaDSA), evaluating a patient's depression level on multiple criteria by incorporating an assessment-aided response generation.

Abstract

With the advancement of chatbots and the growing demand for automatic depression detection, identifying depression in patient conversations has gained more attention. However, prior methods often assess depression in a binary way or only a single score without diverse feedback and lack focus on enhancing dialogue responses. In this paper, we present a novel task of multi-aspect depression severity assessment via an inductive dialogue system (MaDSA), evaluating a patient's depression level on multiple criteria by incorporating an assessment-aided response generation. Further, we propose a foundational system for MaDSA, which induces psychological dialogue responses with an auxiliary emotion classification task within a hierarchical severity assessment structure. We synthesize the conversational dataset annotated with eight aspects of depression severity alongside emotion labels, proven robust via human evaluations. Experimental results show potential for our preliminary work on MaDSA.

Multi-aspect Depression Severity Assessment via Inductive Dialogue System

TL;DR

A novel task of multi-aspect depression severity assessment via an inductive dialogue system (MaDSA), evaluating a patient's depression level on multiple criteria by incorporating an assessment-aided response generation.

Abstract

With the advancement of chatbots and the growing demand for automatic depression detection, identifying depression in patient conversations has gained more attention. However, prior methods often assess depression in a binary way or only a single score without diverse feedback and lack focus on enhancing dialogue responses. In this paper, we present a novel task of multi-aspect depression severity assessment via an inductive dialogue system (MaDSA), evaluating a patient's depression level on multiple criteria by incorporating an assessment-aided response generation. Further, we propose a foundational system for MaDSA, which induces psychological dialogue responses with an auxiliary emotion classification task within a hierarchical severity assessment structure. We synthesize the conversational dataset annotated with eight aspects of depression severity alongside emotion labels, proven robust via human evaluations. Experimental results show potential for our preliminary work on MaDSA.

Paper Structure

This paper contains 15 sections, 6 equations, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. MaDSA
  4. Dataset Synthesis
  5. Inductive Dialogue System
  6. Depression Severity Assessment
  7. Experiments
  8. Settings
  9. Evaluation Metrics
  10. Results
  11. Main Results
  12. Effects on Response Generation
  13. Human Evaluation for Synthetic Data
  14. Depression Detection on DAIC-WOZ
  15. Conclusion

Figures (3)

  • Figure 1: Example of the synthetic data. The yellow represents self-esteem items and the pink represents mood items.
  • Figure 2: The left is an inductive dialogue system, and the right is a severity assessment system trained on a frozen encoder of the dialogue system.
  • Figure 3: (a) The asterisk shows the Spearman’s correlation between human annotation and our data label, while the bar indicates the QWK scores of MaDSA measured with human labels and synthetic labels respectively. (b) Box plot of the distribution of naturalness scores evaluated by professionals.