CMATH: Cross-Modality Augmented Transformer with Hierarchical Variational Distillation for Multimodal Emotion Recognition in Conversation

TL;DR

This paper proposes a novel Cross-Modality Augmented Transformer with Hierarchical Variational Distillation, called CMATH, which consists of two major components, i.e., Multimodal Interaction Fusion and Hierarchical Variational Distillation.

Abstract

Multimodal emotion recognition in conversation (MER) aims to accurately identify emotions in conversational utterances by integrating multimodal information. Previous methods usually treat multimodal information as equal quality and employ symmetric architectures to conduct multimodal fusion. However, in reality, the quality of different modalities usually varies considerably, and utilizing a symmetric architecture is difficult to accurately recognize conversational emotions when dealing with uneven modal information. Furthermore, fusing multi-modality information in a single granularity may fail to adequately integrate modal information, exacerbating the inaccuracy in emotion recognition. In this paper, we propose a novel Cross-Modality Augmented Transformer with Hierarchical Variational Distillation, called CMATH, which consists of two major components, i.e., Multimodal Interaction Fusion and Hierarchical Variational Distillation. The former is comprised of two submodules, including Modality Reconstruction and Cross-Modality Augmented Transformer (CMA-Transformer), where Modality Reconstruction focuses on obtaining high-quality compressed representation of each modality, and CMA-Transformer adopts an asymmetric fusion strategy which treats one modality as the central modality and takes others as auxiliary modalities. The latter first designs a variational fusion network to fuse the fine-grained representations learned by CMA- Transformer into a coarse-grained representations. Then, it introduces a hierarchical distillation framework to maintain the consistency between modality representations with different granularities. Experiments on the IEMOCAP and MELD datasets demonstrate that our proposed model outperforms previous state-of-the-art baselines. Implementation codes can be available at https://github.com/ cjw-MER/CMATH.

Figures (7)

• Figure 1: An example of a multimodal conversation scenario, in which the conversational contents are from textual, audio and visual modalities.
• Figure 2: Overview of our proposed model CMATH, which consists of two modules: (1) Multimodal Interaction Fusion, which uses interaction information to improve modality quality; (2) Hierarchical Variational Distillation, which further boosts the fusion effectiveness between modalities.
• Figure 3: The t-SNE visualizations of learned embeddings by CMATH and MultiDAG on the IEMOCAP dataset.
• Figure 4: Performance of the proposed model with different dimension size on the two datasets.
• Figure 5: Performance of the proposed model with different conversation lengths on the two datasets.