MISA: Modality-Invariant and -Specific Representations for Multimodal Sentiment Analysis

TL;DR

This paper introduces MISA, a novel framework for multimodal sentiment analysis that learns both shared and modality-specific representations to improve fusion and prediction accuracy across multiple datasets.

Contribution

MISA proposes a dual subspace approach to learn modality-invariant and -specific features, enhancing multimodal fusion for sentiment analysis.

Findings

Significant improvements on MOSI and MOSEI benchmarks.

Outperforms state-of-the-art models in sentiment prediction.

Effective in humor detection on UR_FUNNY dataset.

Abstract

Multimodal Sentiment Analysis is an active area of research that leverages multimodal signals for affective understanding of user-generated videos. The predominant approach, addressing this task, has been to develop sophisticated fusion techniques. However, the heterogeneous nature of the signals creates distributional modality gaps that pose significant challenges. In this paper, we aim to learn effective modality representations to aid the process of fusion. We propose a novel framework, MISA, which projects each modality to two distinct subspaces. The first subspace is modality-invariant, where the representations across modalities learn their commonalities and reduce the modality gap. The second subspace is modality-specific, which is private to each modality and captures their characteristic features. These representations provide a holistic view of the multimodal data, which is used for fusion that leads to task predictions. Our experiments on popular sentiment analysis benchmarks, MOSI and MOSEI, demonstrate significant gains over state-of-the-art models. We also consider the task of Multimodal Humor Detection and experiment on the recently proposed UR_FUNNY dataset. Here too, our model fares better than strong baselines, establishing MISA as a useful multimodal framework.