PIXELMOD: Improving Soft Moderation of Visual Misleading Information on Twitter
This addresses the challenge of moderating visual misinformation on social media platforms like Twitter, though it appears incremental as it builds on existing image similarity approaches.
The paper tackles the problem of identifying image-based misinformation at scale on Twitter by presenting PIXELMOD, a system that uses perceptual hashes, vector databases, and OCR to efficiently detect visually misleading images for soft moderation, achieving 0.99% false detection and 2.06% false negatives on a dataset from the 2020 US Presidential Election.
Images are a powerful and immediate vehicle to carry misleading or outright false messages, yet identifying image-based misinformation at scale poses unique challenges. In this paper, we present PIXELMOD, a system that leverages perceptual hashes, vector databases, and optical character recognition (OCR) to efficiently identify images that are candidates to receive soft moderation labels on Twitter. We show that PIXELMOD outperforms existing image similarity approaches when applied to soft moderation, with negligible performance overhead. We then test PIXELMOD on a dataset of tweets surrounding the 2020 US Presidential Election, and find that it is able to identify visually misleading images that are candidates for soft moderation with 0.99% false detection and 2.06% false negatives.