Open-Domain, Content-based, Multi-modal Fact-checking of Out-of-Context Images via Online Resources
release_hlacnjlqs5devhmxgkt6aheyoe
by
Sahar Abdelnabi, Rakibul Hasan, Mario Fritz
2021
Abstract
Misinformation is now a major problem due to its potential high risks to our
core democratic and societal values and orders. Out-of-context misinformation
is one of the easiest and effective ways used by adversaries to spread viral
false stories. In this threat, a real image is re-purposed to support other
narratives by misrepresenting its context and/or elements. The internet is
being used as the go-to way to verify information using different sources and
modalities. Our goal is an inspectable method that automates this
time-consuming and reasoning-intensive process by fact-checking the
image-caption pairing using Web evidence. To integrate evidence and cues from
both modalities, we introduce the concept of 'multi-modal cycle-consistency
check'; starting from the image/caption, we gather textual/visual evidence,
which will be compared against the other paired caption/image, respectively.
Moreover, we propose a novel architecture, Consistency-Checking Network (CCN),
that mimics the layered human reasoning across the same and different
modalities: the caption vs. textual evidence, the image vs. visual evidence,
and the image vs. caption. Our work offers the first step and benchmark for
open-domain, content-based, multi-modal fact-checking, and significantly
outperforms previous baselines that did not leverage external evidence.
In text/plain
format
Archived Files and Locations
application/pdf 17.2 MB
file_2uorya6j3bdhfojiremfzkgrde
|
arxiv.org (repository) web.archive.org (webarchive) |
2112.00061v1
access all versions, variants, and formats of this works (eg, pre-prints)