Fitting Different Interactive Information: Joint Classification of Emotion and Intention

TL;DR

This paper presents a novel approach for low-resource multimodal emotion and intention recognition, utilizing pseudo-labeling and task interaction to improve performance, leading to first place in the ICASSP MEIJU@2025 competition.

Contribution

It introduces a joint classification framework that leverages unlabeled data and task interaction to enhance emotion and intention recognition in low-resource settings.

Findings

Achieved a test score of 0.5532, winning the competition.

Effective use of pseudo-labeling improves recognition accuracy.

Task interaction promotes mutual enhancement of emotion and intention detection.

Abstract

This paper is the first-place solution for ICASSP MEIJU@2025 Track I, which focuses on low-resource multimodal emotion and intention recognition. How to effectively utilize a large amount of unlabeled data, while ensuring the mutual promotion of different difficulty levels tasks in the interaction stage, these two points become the key to the competition. In this paper, pseudo-label labeling is carried out on the model trained with labeled data, and samples with high confidence and their labels are selected to alleviate the problem of low resources. At the same time, the characteristic of easy represented ability of intention recognition found in the experiment is used to make mutually promote with emotion recognition under different attention heads, and higher performance of intention recognition is achieved through fusion. Finally, under the refined processing data, we achieve the score of 0.5532 in the Test set, and win the championship of the track.