Summary of Msp-podcast Ser Challenge 2024: L’antenne Du Ventoux Multimodal Self-supervised Learning For Speech Emotion Recognition, by Jarod Duret (lia) et al.
MSP-Podcast SER Challenge 2024: L’antenne du Ventoux Multimodal Self-Supervised Learning for Speech Emotion Recognition
by Jarod Duret, Mickael Rouvier, Yannick Estève
First submitted to arxiv on: 8 Jul 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Sound (cs.SD); Audio and Speech Processing (eess.AS)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper presents a submission to the 2024 MSP-Podcast Speech Emotion Recognition (SER) Challenge, which consists of two tasks: Categorical Emotion Recognition and Emotional Attribute Prediction. The authors focus on Task 1, categorizing eight emotional states using data from the MSP-Podcast dataset with an ensemble of models and a Support Vector Machine (SVM) classifier. The models are trained using various strategies, including Self-Supervised Learning (SSL) fine-tuning across different modalities: speech alone, text alone, and a combined approach. This paper’s contribution aims to enhance the system’s ability to accurately classify emotional states, achieving an F1-macro of 0.35% on the development set. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about improving computers’ ability to recognize emotions in audio recordings. It’s part of a bigger competition that challenges teams to develop better systems for recognizing emotions. The authors focus on one specific task: categorizing eight different emotional states, like happy or sad. They use a combination of computer models and training methods to make their system better. By combining different types of data, like just speech or just text, they can improve the accuracy of their predictions. |
Keywords
» Artificial intelligence » Fine tuning » Self supervised » Support vector machine
