# MA-DLE: Speech-based Automatic Depression Level Estimation via Memory Augmentation

arXiv: https://arxiv.org/abs/2606.11197 · HTML: https://mlsanity.com/paper/2606.11197/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 82 references that are in the map.
- First posted on arXiv 20 Apr 2026.

## Builds on (most cited first, 21 of 21)

- [InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory](https://mlsanity.com/paper/2402.04617/) (2024)
- [Emformer: Efficient Memory Transformer Based Acoustic Model For Low Latency Streaming Speech Recognition](https://mlsanity.com/paper/2010.10759/) (2020)
- [AVEC 2019 Workshop and Challenge: State-of-Mind, Detecting Depression with AI, and Cross-Cultural Affect Recognition](https://mlsanity.com/paper/1907.11510/) (2019)
- [A Distractor-Aware Memory for Visual Object Tracking with SAM2](https://mlsanity.com/paper/2411.17576/) (2024)
- [Video Prediction Recalling Long-term Motion Context via Memory Alignment Learning](https://mlsanity.com/paper/2104.00924/) (2021)
- [CubeMLP: An MLP-based Model for Multimodal Sentiment Analysis and Depression Estimation](https://mlsanity.com/paper/2207.14087/) (2022)
- [Deep Learning for Depression Recognition with Audiovisual Cues: A Review](https://mlsanity.com/paper/2106.00610/) (2021)
- [DEPA: Self-Supervised Audio Embedding for Depression Detection](https://mlsanity.com/paper/1910.13028/) (2019)
- [Multi-modal Depression Estimation based on Sub-attentional Fusion](https://mlsanity.com/paper/2207.06180/) (2022)
- [Context-Enhanced Memory-Refined Transformer for Online Action Detection](https://mlsanity.com/paper/2503.18359/) (2025)
- [Multimodal Emotion Recognition in Conversations: A Survey of Methods, Trends, Challenges and Prospects](https://mlsanity.com/paper/2505.20511/) (2025)
- [Two-stage Temporal Modelling Framework for Video-based Depression Recognition using Graph Representation](https://mlsanity.com/paper/2111.15266/) (2021)
- [Depression Severity Estimation from Multiple Modalities](https://mlsanity.com/paper/1711.06095/) (2017)
- [Face-Driven Zero-Shot Voice Conversion with Memory-based Face-Voice Alignment](https://mlsanity.com/paper/2309.09470/) (2023)
- [Depression Diagnosis and Analysis via Multimodal Multi-order Factor Fusion](https://mlsanity.com/paper/2301.00254/) (2022)
- [NUC-Net: Non-uniform Cylindrical Partition Network for Efficient LiDAR Semantic Segmentation](https://mlsanity.com/paper/2505.24634/) (2025)
- [AdaSFormer: Adaptive Serialized Transformers for Monocular Semantic Scene Completion from Indoor Environments](https://mlsanity.com/paper/2603.25494/) (2026)
- [Attention-Based Acoustic Feature Fusion Network for Depression Detection](https://mlsanity.com/paper/2308.12478/) (2023)
- [Conformal Depression Prediction](https://mlsanity.com/paper/2405.18723/) (2024)
- [Investigating Acoustic-Textual Emotional Inconsistency Information for Automatic Depression Detection](https://mlsanity.com/paper/2412.18614/) (2024)
- [Efficient Long Speech Sequence Modelling for Time-Domain Depression Level Estimation](https://mlsanity.com/paper/2501.02512/) (2025)

## Abstract (arXiv)

Speech-based automatic estimation of depression levels is essential for enabling early detection and timely intervention, particularly in resource-constrained mental health settings. In recent years, deep learning has demonstrated impressive success across various domains, including affective computing and mental health assessment. Most existing approaches rely on RNN-based architectures (such as LSTM and GRU) to model temporal information for depression estimation. However, the extracted features often emphasize only a few adjacent speech segments, limiting their ability to capture long-range dependencies. To overcome this limitation, we introduce a memory-based feature augmentation method that enhances the representational capacity of GRU-extracted features. Rather than indiscriminately incorporating historical data, our memory bank is designed to selectively integrate two types of components in order to reduce redundancy and irrelevance: (1) historical temporal features that closely resemble the current GRU output, offering complementary contextual information; and (2) dynamic memory features identified based on feature variability, which capture behavioral and emotional fluctuations indicative of depressive symptoms. To effectively fuse the memory-augmented features with GRU outputs, we further design a Hierarchical Attention Fusion (HAF) module. Our method is evaluated on the widely used DAIC-WOZ and E-DAIC datasets, achieving state-of-the-art performance.