MMOOC: A Multimodal Misinformation Dataset for Out-of-Context News Analysis

Xu, Q; Du, H; Chen, H; Liu, B; Yu, X

MMOOC: A Multimodal Misinformation Dataset for Out-of-Context News Analysis

Xu, Q Du, H Chen, H Liu, B

Yu, X

Permalink

Publisher:: Springer Nature
Publication Type:: Chapter
Citation:: Information Security and Privacy, 2024, 14897 LNCS, pp. 444-459
Issue Date:: 2024-01-01

Embargoed

	Filename	Description	Size
	MMOOC: A Multimodal Misinformation Dataset for Out-of-Context News Analysis.pdf	Accepted version	1.19 MB	Adobe PDF	View/Open

Copyright Clearance Process

Recently Added
In Progress
Embargoed
Open Access

This item is currently unavailable due to the publisher's embargo.

The embargo period expires on 15 Jul 2026

Full metadata record

Field	Value	Language
dc.contributor.author	Xu, Q
dc.contributor.author	Du, H
dc.contributor.author	Chen, H
dc.contributor.author	Liu, B https://orcid.org/0000-0002-3603-6617
dc.contributor.author	Yu, X https://orcid.org/0000-0002-0269-5649
dc.date.accessioned	2024-08-14T00:59:37Z
dc.date.available	2024-08-14T00:59:37Z
dc.date.issued	2024-01-01
dc.identifier.citation	Information Security and Privacy, 2024, 14897 LNCS, pp. 444-459
dc.identifier.isbn	9789819751006
dc.identifier.uri	http://hdl.handle.net/10453/180403
dc.description.abstract	Misinformation in the news media poses a significant challenge, particularly with the rise of manipulating real news into misleading contexts, i.e., Out-of-Context (OOC) media. Existing datasets for studying OOC media only have at most two modalities with limited source and topic scope. These datasets undermine the effectiveness of models trained to identify real-world misinformation. In this paper, we first introduce a comprehensive OOC media dataset compiled from various sources, dubbed the Multimodal Misinformation Dataset for Out-of-Context News Analysis (MMOOC). We collect 91K authentic multimodal news from 60 influential news outlets around the world, such as ABC News and BBC News. Then, we produce 364k OOC fabricated news data by recombining the authentic ones. Furthermore, we propose an MMOOC-Checker to check the OOC media by leveraging not only the semantic consistency among different modalities but also temporal consistency between the first released dates of the modalities. To be specific, we develop an internal OOC-checker to examine the semantic consistency between modalities. Meanwhile, we design an external OOC-checker to utilize the temporal closeness between the news in different modalities. Experiments on MMOOC demonstrate the effectiveness of the MMOOC-Checker. The dataset will be released soon.
dc.language	en
dc.publisher	Springer Nature
dc.relation	http://purl.org/au-research/grants/arc/DP230100246
dc.relation	http://purl.org/au-research/grants/arc/LP220200808
dc.relation.ispartof	Information Security and Privacy
dc.relation.ispartofseries	Lecture Notes in Computer Science
dc.relation.isbasedon	10.1007/978-981-97-5101-3_24
dc.rights	info:eu-repo/semantics/embargoedAccess
dc.subject.classification	Artificial Intelligence & Image Processing
dc.subject.classification	46 Information and computing sciences
dc.title	MMOOC: A Multimodal Misinformation Dataset for Out-of-Context News Analysis
dc.type	Chapter
utslib.citation.volume	14897 LNCS
pubs.organisational-group	University of Technology Sydney
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology
pubs.organisational-group	University of Technology Sydney/Strength - AAII - Australian Artificial Intelligence Institute
pubs.organisational-group	University of Technology Sydney/Faculty of Engineering and Information Technology/School of Computer Science
pubs.organisational-group	University of Technology Sydney/Strength - CCSP - Centre for Cyber Security and Privacy
utslib.copyright.status	embargoed	*
utslib.copyright.embargo	2026-07-15T00:00:00+1000Z
dc.date.updated	2024-08-14T00:59:35Z
pubs.publication-status	Published
pubs.volume	14897 LNCS

Abstract:

Misinformation in the news media poses a significant challenge, particularly with the rise of manipulating real news into misleading contexts, i.e., Out-of-Context (OOC) media. Existing datasets for studying OOC media only have at most two modalities with limited source and topic scope. These datasets undermine the effectiveness of models trained to identify real-world misinformation. In this paper, we first introduce a comprehensive OOC media dataset compiled from various sources, dubbed the Multimodal Misinformation Dataset for Out-of-Context News Analysis (MMOOC). We collect 91K authentic multimodal news from 60 influential news outlets around the world, such as ABC News and BBC News. Then, we produce 364k OOC fabricated news data by recombining the authentic ones. Furthermore, we propose an MMOOC-Checker to check the OOC media by leveraging not only the semantic consistency among different modalities but also temporal consistency between the first released dates of the modalities. To be specific, we develop an internal OOC-checker to examine the semantic consistency between modalities. Meanwhile, we design an external OOC-checker to utilize the temporal closeness between the news in different modalities. Experiments on MMOOC demonstrate the effectiveness of the MMOOC-Checker. The dataset will be released soon.

Please use this identifier to cite or link to this item:

http://hdl.handle.net/10453/180403