Memes in the Wild: Assessing the Generalizability of the Hateful Memes Challenge Dataset

TL;DR

This paper evaluates the generalizability of hateful meme detection models trained on a Facebook dataset by testing them on real-world memes from Pinterest, highlighting challenges like OCR noise and meme diversity.

Contribution

It provides an empirical assessment of how well existing models perform on memes in the wild, revealing key differences and limitations.

Findings

OCR noise reduces model performance

Memes in the wild are more diverse

Current benchmarks may not reflect real-world scenarios

Abstract

Hateful memes pose a unique challenge for current machine learning systems because their message is derived from both text- and visual-modalities. To this effect, Facebook released the Hateful Memes Challenge, a dataset of memes with pre-extracted text captions, but it is unclear whether these synthetic examples generalize to `memes in the wild'. In this paper, we collect hateful and non-hateful memes from Pinterest to evaluate out-of-sample performance on models pre-trained on the Facebook dataset. We find that memes in the wild differ in two key aspects: 1) Captions must be extracted via OCR, injecting noise and diminishing performance of multimodal models, and 2) Memes are more diverse than `traditional memes', including screenshots of conversations or text on a plain background. This paper thus serves as a reality check for the current benchmark of hateful meme detection and its applicability for detecting real world hate.