# Uncovering Latent Depression Severity for Binary Depression Detection via Advantage-weighting Ranking

arXiv: https://arxiv.org/abs/2607.05901 · HTML: https://mlsanity.com/paper/2607.05901/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 27 references that are in the map.
- First posted on arXiv 7 Jul 2026.

## Builds on (most cited first, 7 of 7)

- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verification](https://mlsanity.com/paper/2005.07143/) (2020)
- [EPIC-Fusion: Audio-Visual Temporal Binding for Egocentric Action Recognition](https://mlsanity.com/paper/1908.08498/) (2019)
- [DepMamba: Progressive Fusion Mamba for Multimodal Depression Detection](https://mlsanity.com/paper/2409.15936/) (2024)
- [LMVD: A Large-Scale Multimodal Vlog Dataset for Depression Detection in the Wild](https://mlsanity.com/paper/2407.00024/) (2024)
- [MDD-Net: Multimodal Depression Detection through Mutual Transformer](https://mlsanity.com/paper/2508.08093/) (2025)
- [CAF-Mamba: Mamba-Based Cross-Modal Adaptive Attention Fusion for Multimodal Depression Detection](https://mlsanity.com/paper/2601.21648/) (2026)

## Abstract (arXiv)

Automatic depression detection using audio-visual data faces significant challenges, particularly in disentangling overlapping feature distributions and establishing robust decision boundaries. To address this, we propose a fine-grained multimodal framework featuring a temporal encoder and a mutual transformer to facilitate deep cross-modal fusion. Our core contribution is the Binary Advantage-weighting Ranking Loss, which optimizes the latent space distribution through two complementary mechanisms: Advantage-weighted Separation, which mines hard pairs by computing a pairwise prediction difference matrix and dynamically weighting them based on their difficulty; and Advantage-weighted Compactness, which minimizes intra-class variance to force features to cluster around their respective class centers. Extensive experiments on D-vlog and LMVD demonstrate that our model reconstructs the latent ordinal structure by prioritizing hard pairs, thereby achieving state-of-the-art performance.