Unimodal Intermediate Training for Multimodal Meme Sentiment Classification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hazman, Muzhaffar, McKeever, Susan, Griffith, Josephine
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908480417300480
author Hazman, Muzhaffar
McKeever, Susan
Griffith, Josephine
author_facet Hazman, Muzhaffar
McKeever, Susan
Griffith, Josephine
contents Internet Memes remain a challenging form of user-generated content for automated sentiment classification. The availability of labelled memes is a barrier to developing sentiment classifiers of multimodal memes. To address the shortage of labelled memes, we propose to supplement the training of a multimodal meme classifier with unimodal (image-only and text-only) data. In this work, we present a novel variant of supervised intermediate training that uses relatively abundant sentiment-labelled unimodal data. Our results show a statistically significant performance improvement from the incorporation of unimodal text data. Furthermore, we show that the training set of labelled memes can be reduced by 40% without reducing the performance of the downstream model.
format Preprint
id arxiv_https___arxiv_org_abs_2308_00528
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Unimodal Intermediate Training for Multimodal Meme Sentiment Classification
Hazman, Muzhaffar
McKeever, Susan
Griffith, Josephine
Computation and Language
Internet Memes remain a challenging form of user-generated content for automated sentiment classification. The availability of labelled memes is a barrier to developing sentiment classifiers of multimodal memes. To address the shortage of labelled memes, we propose to supplement the training of a multimodal meme classifier with unimodal (image-only and text-only) data. In this work, we present a novel variant of supervised intermediate training that uses relatively abundant sentiment-labelled unimodal data. Our results show a statistically significant performance improvement from the incorporation of unimodal text data. Furthermore, we show that the training set of labelled memes can be reduced by 40% without reducing the performance of the downstream model.
title Unimodal Intermediate Training for Multimodal Meme Sentiment Classification
topic Computation and Language
url https://arxiv.org/abs/2308.00528