M2DF: Multi-grained Multi-curriculum Denoising Framework for Multimodal Aspect-based Sentiment Analysis

TL;DR

This paper introduces M2DF, a curriculum learning-based framework that reduces the negative impact of noisy images in multimodal aspect-based sentiment analysis without data modification, improving performance over existing methods.

Contribution

The paper proposes a novel multi-grained multi-curriculum denoising framework (M2DF) that enhances MABSA by effectively mitigating noise from irrelevant images during training.

Findings

Outperforms state-of-the-art on three MABSA sub-tasks

Consistently improves model robustness to noisy images

Effective without data filtering or thresholding

Abstract

Multimodal Aspect-based Sentiment Analysis (MABSA) is a fine-grained Sentiment Analysis task, which has attracted growing research interests recently. Existing work mainly utilizes image information to improve the performance of MABSA task. However, most of the studies overestimate the importance of images since there are many noise images unrelated to the text in the dataset, which will have a negative impact on model learning. Although some work attempts to filter low-quality noise images by setting thresholds, relying on thresholds will inevitably filter out a lot of useful image information. Therefore, in this work, we focus on whether the negative impact of noisy images can be reduced without modifying the data. To achieve this goal, we borrow the idea of Curriculum Learning and propose a Multi-grained Multi-curriculum Denoising Framework (M2DF), which can achieve denoising by adjusting the order of training data. Extensive experimental results show that our framework consistently outperforms state-of-the-art work on three sub-tasks of MABSA.