MERID: Multimodal Exploration via Recursive Self-Improvement Agents for Major Depression Analysis

2026-09-28

AI Summary

This study addresses the challenges of multimodal data heterogeneity, difficult prediction pipeline customization, and the lack of autonomous iteration in major depressive disorder detection by proposing the MERID framework. Built upon an experience-driven recursive self-improving agent, MERID introduces a novel integration of grounded state construction, coupled pipeline exploration, and evidence-guided evolution. By leveraging multi-agent systems and uncertainty quantification, it achieves knowledge inheritance of validated gains and automatic pipeline optimization under few-shot conditions. Experimental results demonstrate that MERID surpasses existing baselines to achieve state-of-the-art performance across multiple depression detection benchmarks, while revealing the critical diagnostic value of acoustic and linguistic cues.

Abstract

Major depressive disorder (MDD) severely impacts daily activities and quality of life. Detecting MDD involves multimodal data, such as interview recordings and sensor measurements. This is particularly challenging, as these heterogeneous modalities often demand distinct, customized prediction pipelines. Existing efforts to address this challenge have explored both manually engineered multimodal architectures and agent-assisted pipeline development. Despite their progress, it remains challenging to autonomously revise pipelines based on experimental feedback and carry verified improvements forward into subsequent designs. To this end, we propose Multimodal Exploration via Recursive Self-Improvement Agents for Major Depression Analysis (MERID). The framework develops depression pipelines through experience-based recursive self-improvement (RSI). Grounded State Construction (GSC) grounds experience by aligning multimodal records with subject-level depression targets. Coupled Pipeline Exploration (CPE) jointly modifies representations, fusion, and predictors to build successor pipelines for classification and severity estimation. Evidence-Guided Evolution (EGE) guides revisions through feedback and verifies gains under uncertainty in small depression cohorts before inheritance. Extensive experiments on depression benchmarks show that MERID achieves the best results on multiple tasks compared with multimodal and agent-based baselines. Further analysis highlights the value of acoustic and linguistic cues for depression detection. Our code is available at https://github.com/DiscoAILab/MERID

Problem

• Major Depressive Disorder
• Multimodal Data
• Pipeline Optimization
• Recursive Self-Improvement

Innovation

• Recursive Self-Improvement
• Multimodal Exploration
• Major Depressive Disorder
• Pipeline Evolution
• AI Agents