Mapping Visual Themes among Authentic and Coordinated Memes

TL;DR

This paper uses a self-supervised vision model to analyze and compare visual themes in authentic versus coordinated memes, revealing significant thematic overlaps and distinguishing features with high accuracy.

Contribution

It introduces a novel application of DeepCluster embeddings and clustering to differentiate authentic and state-sponsored memes based on visual themes.

Findings

Coordinated memes focus more on celebrities, military, and gender themes.

Authentic memes feature more comics and movie characters.

Logistic regression achieves 84% accuracy in identifying IRA memes.

Abstract

What distinguishes authentic memes from those created by state actors? I utilize a self-supervised vision model, DeepCluster (Caron et al. 2019), to learn low-dimensional visual embeddings of memes and apply K-means to jointly cluster authentic and coordinated memes without additional inputs. I find that authentic and coordinated memes share a large fraction of visual themes but with varying degrees. Coordinated memes from Russian IRA accounts promote more themes around celebrities, quotes, screenshots, military, and gender. Authentic Reddit memes include more themes with comics and movie characters. A simple logistic regression on the low-dimensional embeddings can discern IRA memes from Reddit memes with an out-sample testing accuracy of 0.84.