Audio-Guided Fusion Techniques for Multimodal Emotion Analysis

TL;DR

This paper introduces an audio-guided fusion approach for multimodal emotion analysis, combining fine-tuned feature extractors, a novel transformer fusion mechanism, and self-supervised learning to improve sentiment classification accuracy.

Contribution

It presents a new Audio-Guided Transformer fusion mechanism and a semi-supervised learning strategy for multimodal emotion analysis, achieving competitive results in MER2024.

Findings

Achieved third place in MER-SEMI track

Effective fusion of audio, video, and text features

Improved sentiment classification performance

Abstract

In this paper, we propose a solution for the semi-supervised learning track (MER-SEMI) in MER2024. First, in order to enhance the performance of the feature extractor on sentiment classification tasks,we fine-tuned video and text feature extractors, specifically CLIP-vit-large and Baichuan-13B, using labeled data. This approach effectively preserves the original emotional information conveyed in the videos. Second, we propose an Audio-Guided Transformer (AGT) fusion mechanism, which leverages the robustness of Hubert-large, showing superior effectiveness in fusing both inter-channel and intra-channel information. Third, To enhance the accuracy of the model, we iteratively apply self-supervised learning by using high-confidence unlabeled data as pseudo-labels. Finally, through black-box probing, we discovered an imbalanced data distribution between the training and test sets. Therefore, We adopt a prior-knowledge-based voting mechanism. The results demonstrate the effectiveness of our strategy, ultimately earning us third place in the MER-SEMI track.