English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Modality representation learning is an important problem for multimodalsentiment analysis (MSA), since the highly distinguishable representations cancontribute to improving the analysis effect. Previous works of MSA have usuallyfocused on multimodal fusion strategies, and the deep study of modalrepresentation learning was given less attention. Recently, contrastivelearning has been confirmed effective at endowing the learned representationwith stronger discriminate ability. Inspired by this, we explore theimprovement approaches of modality representation with contrastive learning inthis study. To this end, we devise a three-stages framework with multi-viewcontrastive learning to refine representations for the specific objectives. Atthe first stage, for the improvement of unimodal representations, we employ thesupervised contrastive learning to pull samples within the same class togetherwhile the other samples are pushed apart. At the second stage, aself-supervised contrastive learning is designed for the improvement of thedistilled unimodal representations after cross-modal interaction. At last, weleverage again the supervised contrastive learning to enhance the fusedmultimodal representation. After all the contrast trainings, we next achievethe classification task based on frozen representations. We conduct experimentson three open datasets, and results show the advance of our model.

Improving the Modality Representation with Multi-View Contrastive Learning for Multimodal Sentiment Analysis