A Multimodal Framework for the Detection of Hateful Memes

Abstract

An increasingly common expression of online hate speech is multimodal in nature and comes in the form of memes. Designing systems to automatically detect hateful content is of paramount importance if we are to mitigate its undesirable effects on the society at large. The detection of multimodal hate speech is an intrinsically difficult and open problem: memes convey a message using both images and text and, hence, require multimodal reasoning and joint visual and language understanding. In this work, we seek to advance this line of research and develop a multimodal framework for the detection of hateful memes. We improve the performance of existing multimodal approaches beyond simple fine-tuning and, among others, show the effectiveness of upsampling of contrastive examples to encourage multimodality and ensemble learning based on cross-validation to improve robustness. We furthermore analyze model misclassifications and discuss a number of hypothesis-driven augmentations and their effects on performance, presenting important implications for future research in the field. Our best approach comprises an ensemble of UNITER-based models and achieves an AUROC score of 80.53, placing us 4th on phase 2 of the 2020 Hateful Memes Challenge organized by Facebook.

References in corpus (9)

• Adam: A Method for Stochastic Optimization
• VisualBERT: A Simple and Performant Baseline for Vision and Language
• VQA: Visual Question Answering
• LXMERT: Learning Cross-Modality Encoder Representations from Transformers
• Supervised Multimodal Bitransformers for Classifying Images and Text
• The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes
• ERNIE-ViL: Knowledge Enhanced Vision-Language Representations Through Scene Graph
• Prediction of Cyberbullying Incidents on the Instagram Social Network
• Tackling Online Abuse: A Survey of Automated Abuse Detection Methods

Cited by in corpus (10)

• Learning Transferable Visual Models From Natural Language Supervision
• Disentangling Hate in Online Memes
• SoK: Content Moderation in Social Media, from Guidelines to Enforcement, and Research to Practice
• PromptMTopic: Unsupervised Multimodal Topic Modeling of Memes using Large Language Models
• Toxic Memes: A Survey of Computational Perspectives on the Detection and Explanation of Meme Toxicities
• HateProof: Are Hateful Meme Detection Systems really Robust?
• Enhance Multimodal Model Performance with Data Augmentation: Facebook Hateful Meme Challenge Solution
• MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets
• Detecting Propaganda Techniques in Memes
• Detecting Harmful Memes and Their Targets