Multimodal Learning using Optimal Transport for Sarcasm and Humor Detection

TL;DR

This paper introduces MuLOT, a novel multimodal learning system that employs optimal transport and self-attention to improve sarcasm and humor detection in resource-limited settings across various multimedia datasets.

Contribution

The paper presents MuLOT, a new multimodal learning framework combining optimal transport and self-attention for sarcasm and humor detection with limited training data.

Findings

Achieved 2.1%, 1.54%, and 2.34% accuracy improvements over state-of-the-art methods.

Effectively captures intra- and inter-modal dependencies using self-attention and multimodal fusion.

Demonstrates robustness across three benchmark datasets for multimodal sarcasm and humor detection.

Abstract

Multimodal learning is an emerging yet challenging research area. In this paper, we deal with multimodal sarcasm and humor detection from conversational videos and image-text pairs. Being a fleeting action, which is reflected across the modalities, sarcasm detection is challenging since large datasets are not available for this task in the literature. Therefore, we primarily focus on resource-constrained training, where the number of training samples is limited. To this end, we propose a novel multimodal learning system, MuLOT (Multimodal Learning using Optimal Transport), which utilizes self-attention to exploit intra-modal correspondence and optimal transport for cross-modal correspondence. Finally, the modalities are combined with multimodal attention fusion to capture the inter-dependencies across modalities. We test our approach for multimodal sarcasm and humor detection on three benchmark datasets - MUStARD (video, audio, text), UR-FUNNY (video, audio, text), MST (image, text) and obtain 2.1%, 1.54%, and 2.34% accuracy improvements over state-of-the-art.