Speech-Based Depression Prediction Using Encoder-Weight-Only Transfer Learning and a Large Corpus

TL;DR

This paper presents a transfer learning method using a lightweight encoder for speech-based depression prediction, demonstrating significant performance improvements on a large dataset and highlighting its efficiency and flexibility.

Contribution

It introduces a transfer learning approach that transfers only encoder weights, enabling a simplified and efficient model for depression prediction from speech.

Findings

Up to 27% relative performance gains in binary depression classification

Statistically significant improvements with large datasets

Transfer learning benefits do not depend on source task performance

Abstract

Speech-based algorithms have gained interest for the management of behavioral health conditions such as depression. We explore a speech-based transfer learning approach that uses a lightweight encoder and that transfers only the encoder weights, enabling a simplified run-time model. Our study uses a large data set containing roughly two orders of magnitude more speakers and sessions than used in prior work. The large data set enables reliable estimation of improvement from transfer learning. Results for the prediction of PHQ-8 labels show up to 27% relative performance gains for binary classification; these gains are statistically significant with a p-value close to zero. Improvements were also found for regression. Additionally, the gain from transfer learning does not appear to require strong source task performance. Results suggest that this approach is flexible and offers promise for efficient implementation.