AttentionDep: Domain-Aware Attention for Explainable Depression Severity Assessment

TL;DR

AttentionDep tackles depression severity estimation from social media by fusing hierarchical textual representations (unigrams and bigrams) with a domain-specific mental health knowledge graph via cross-attention, and by predicting severity with an ordinal regression framework. The approach integrates a Wikipedia-derived MHKG, processed with GINE, to produce domain-informed embeddings that enhance clinical relevance and interpretability. Across three Reddit-based datasets, AttentionDep outperforms twelve baselines by more than 5% in graded F1 and provides explainable signals through unigram/bigram attentions and KG-informed features. This work advances trustworthy, domain-aware, explainable AI for mental health assessment from social media and points to future multimodal extensions to further improve robustness and auditability.

Abstract

In today's interconnected society, social media platforms provide a window into individuals' thoughts, emotions, and mental states. This paper explores the use of platforms like Facebook, X (formerly Twitter), and Reddit for depression severity detection. We propose AttentionDep, a domain-aware attention model that drives explainable depression severity estimation by fusing contextual and domain knowledge. Posts are encoded hierarchically using unigrams and bigrams, with attention mechanisms highlighting clinically relevant tokens. Domain knowledge from a curated mental health knowledge graph is incorporated through a cross-attention mechanism, enriching the contextual features. Finally, depression severity is predicted using an ordinal regression framework that respects the clinical-relevance and natural ordering of severity levels. Our experiments demonstrate that AttentionDep outperforms state-of-the-art baselines by over 5% in graded F1 score across datasets, while providing interpretable insights into its predictions. This work advances the development of trustworthy and transparent AI systems for mental health assessment from social media.

Figures (4)

• Figure 1: Proposed model AttentionDep. It predicts depression severity of a post and generates explanation with unigram- and bigram-level attentions and a knowledge graph.
• Figure 2: Sample snapshot of the Mental Health Knowledge Graph (MHKG)
• Figure 3: Effects of hyperparameters on graded F1 score for severity datasets D4 and D3.
• Figure 4: Token-level importance visualisations for representative posts across different depression severity levels. Hierarchical attention highlights unigrams and bigrams most influential in the model’s predictions, providing insight into the decision-making process.