Developing a High-performance Framework for Speech Emotion Recognition in Naturalistic Conditions Challenge for Emotional Attribute Prediction

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lertpetchpun, Thanathai, Feng, Tiantian, Byrd, Dani, Narayanan, Shrikanth
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866913890354331648
author Lertpetchpun, Thanathai
Feng, Tiantian
Byrd, Dani
Narayanan, Shrikanth
author_facet Lertpetchpun, Thanathai
Feng, Tiantian
Byrd, Dani
Narayanan, Shrikanth
contents Speech emotion recognition (SER) in naturalistic conditions presents a significant challenge for the speech processing community. Challenges include disagreement in labeling among annotators and imbalanced data distributions. This paper presents a reproducible framework that achieves superior (top 1) performance in the Emotion Recognition in Naturalistic Conditions Challenge (IS25-SER Challenge) - Task 2, evaluated on the MSP-Podcast dataset. Our system is designed to tackle the aforementioned challenges through multimodal learning, multi-task learning, and imbalanced data handling. Specifically, our best system is trained by adding text embeddings, predicting gender, and including ``Other'' (O) and ``No Agreement'' (X) samples in the training set. Our system's results secured both first and second places in the IS25-SER Challenge, and the top performance was achieved by a simple two-system ensemble.
format Preprint
id arxiv_https___arxiv_org_abs_2506_10930
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Developing a High-performance Framework for Speech Emotion Recognition in Naturalistic Conditions Challenge for Emotional Attribute Prediction
Lertpetchpun, Thanathai
Feng, Tiantian
Byrd, Dani
Narayanan, Shrikanth
Machine Learning
Speech emotion recognition (SER) in naturalistic conditions presents a significant challenge for the speech processing community. Challenges include disagreement in labeling among annotators and imbalanced data distributions. This paper presents a reproducible framework that achieves superior (top 1) performance in the Emotion Recognition in Naturalistic Conditions Challenge (IS25-SER Challenge) - Task 2, evaluated on the MSP-Podcast dataset. Our system is designed to tackle the aforementioned challenges through multimodal learning, multi-task learning, and imbalanced data handling. Specifically, our best system is trained by adding text embeddings, predicting gender, and including ``Other'' (O) and ``No Agreement'' (X) samples in the training set. Our system's results secured both first and second places in the IS25-SER Challenge, and the top performance was achieved by a simple two-system ensemble.
title Developing a High-performance Framework for Speech Emotion Recognition in Naturalistic Conditions Challenge for Emotional Attribute Prediction
topic Machine Learning
url https://arxiv.org/abs/2506.10930