MSP-Podcast SER Challenge 2024: L'antenne du Ventoux Multimodal Self-Supervised Learning for Speech Emotion Recognition

TL;DR

This paper presents a multimodal self-supervised learning approach for speech emotion recognition, using ensemble models and SVM fusion to classify eight emotional states with improved accuracy.

Contribution

It introduces a novel ensemble and fusion strategy with SSL fine-tuning across speech and text modalities for emotion recognition.

Findings

Achieved F1-macro score of 0.35% on development set

Demonstrated effectiveness of joint speech-text training

Enhanced emotion classification accuracy

Abstract

In this work, we detail our submission to the 2024 edition of the MSP-Podcast Speech Emotion Recognition (SER) Challenge. This challenge is divided into two distinct tasks: Categorical Emotion Recognition and Emotional Attribute Prediction. We concentrated our efforts on Task 1, which involves the categorical classification of eight emotional states using data from the MSP-Podcast dataset. Our approach employs an ensemble of models, each trained independently and then fused at the score level using a Support Vector Machine (SVM) classifier. The models were trained using various strategies, including Self-Supervised Learning (SSL) fine-tuning across different modalities: speech alone, text alone, and a combined speech and text approach. This joint training methodology aims to enhance the system's ability to accurately classify emotional states. This joint training methodology aims to enhance the system's ability to accurately classify emotional states. Thus, the system obtained F1-macro of 0.35\% on development set.