Abstract: Highlights•A deep learning-based depression detection method is proposed as an auxiliary diagnostic tool.•A core question subset is constructed to enhance the efficiency of depression assessment.•Hierarchical weighted attention fusion is proposed to fuse depression-aware features adaptively.•Weighted multi-task learning is used for depression assessment and PHQ-8 score prediction.
External IDs:dblp:journals/spic/WangCZWNYY25
Loading