Hate Speech in Pixels: Detection of Offensive Memes towards Automatic Moderation

TL;DR

This paper explores automatic hate speech detection in memes using visual and linguistic features, highlighting the potential of visual information and providing a new dataset and baseline model for future research.

Contribution

It introduces a new dataset of memes and investigates the effectiveness of visual and linguistic modalities for hate speech detection, emphasizing the importance of visual cues.

Findings

Visual modality is more informative than linguistic for hate speech detection in memes.

The model can detect some hate memes but the task remains challenging.

A new dataset of 5,020 memes was created for this research.

Abstract

This work addresses the challenge of hate speech detection in Internet memes, and attempts using visual information to automatically detect hate speech, unlike any previous work of our knowledge. Memes are pixel-based multimedia documents that contain photos or illustrations together with phrases which, when combined, usually adopt a funny meaning. However, hate memes are also used to spread hate through social networks, so their automatic detection would help reduce their harmful societal impact. Our results indicate that the model can learn to detect some of the memes, but that the task is far from being solved with this simple architecture. While previous work focuses on linguistic hate speech, our experiments indicate how the visual modality can be much more informative for hate speech detection than the linguistic one in memes. In our experiments, we built a dataset of 5,020 memes to train and evaluate a multi-layer perceptron over the visual and language representations, whether independently or fused. The source code and mode and models are available https://github.com/imatge-upc/hate-speech-detection .