Using Auxiliary Tasks In Multimodal Fusion Of Wav2vec 2.0 And BERT For Multimodal Emotion Recognition
This work addresses multimodal emotion recognition for applications like human-computer interaction, but it is incremental as it builds on existing pretrained models and fusion techniques.
The paper tackled the challenges of limited data and multimodal fusion in emotion recognition by fine-tuning pretrained models (wav2vec 2.0 and BERT) and using a multi-head attention fusion module with auxiliary tasks, achieving 78.42% WA and 79.71% UA on the IEMOCAP dataset.
The lack of data and the difficulty of multimodal fusion have always been challenges for multimodal emotion recognition (MER). In this paper, we propose to use pretrained models as upstream network, wav2vec 2.0 for audio modality and BERT for text modality, and finetune them in downstream task of MER to cope with the lack of data. For the difficulty of multimodal fusion, we use a K-layer multi-head attention mechanism as a downstream fusion module. Starting from the MER task itself, we design two auxiliary tasks to alleviate the insufficient fusion between modalities and guide the network to capture and align emotion-related features. Compared to the previous state-of-the-art models, we achieve a better performance by 78.42% Weighted Accuracy (WA) and 79.71% Unweighted Accuracy (UA) on the IEMOCAP dataset.