Addressing Data Imbalance in Transformer-Based Multi-Label Emotion Detection with Weighted Loss

TL;DR

This paper introduces a weighted loss function for Transformer models to improve multi-label emotion detection in imbalanced datasets, showing enhanced performance on common classes but limited gains on minority classes.

Contribution

It proposes a simple, dynamic weighted loss approach for Transformer models to address data imbalance in multi-label emotion detection tasks.

Findings

Improves performance on high-frequency emotion classes.

Limited impact on minority emotion classes.

Demonstrates effectiveness and challenges of weighted loss in imbalanced data.

Abstract

This paper explores the application of a simple weighted loss function to Transformer-based models for multi-label emotion detection in SemEval-2025 Shared Task 11. Our approach addresses data imbalance by dynamically adjusting class weights, thereby enhancing performance on minority emotion classes without the computational burden of traditional resampling methods. We evaluate BERT, RoBERTa, and BART on the BRIGHTER dataset, using evaluation metrics such as Micro F1, Macro F1, ROC-AUC, Accuracy, and Jaccard similarity coefficients. The results demonstrate that the weighted loss function improves performance on high-frequency emotion classes but shows limited impact on minority classes. These findings underscore both the effectiveness and the challenges of applying this approach to imbalanced multi-label emotion detection.