Decoding the Underlying Meaning of Multimodal Hateful Memes

Abstract

Recent studies have proposed models that yielded promising performance for the hateful meme classification task. Nevertheless, these proposed models do not generate interpretable explanations that uncover the underlying meaning and support the classification output. A major reason for the lack of explainable hateful meme methods is the absence of a hateful meme dataset that contains ground truth explanations for benchmarking or training. Intuitively, having such explanations can educate and assist content moderators in interpreting and removing flagged hateful memes. This paper address this research gap by introducing Hateful meme with Reasons Dataset (HatReD), which is a new multimodal hateful meme dataset annotated with the underlying hateful contextual reasons. We also define a new conditional generation task that aims to automatically generate underlying reasons to explain hateful memes and establish the baseline performance of state-of-the-art pre-trained language models on this task. We further demonstrate the usefulness of HatReD by analyzing the challenges of the new conditional generation task in explaining memes in seen and unseen domains. The dataset and benchmark models are made available here: https://github.com/Social-AI-Studio/HatRed

Unravelling passive social media use through screenomes

“…While we did not explore this in greater detail, there is likely substantial variation within these two categories as well. Memes could be political in nature, or be about relationships, work/school, or hateful (Hee et al, 2023). In other words, even though a clustering of social media content into genres and topics is possible, our research illustrates that passive SMU is characterized by truly significant levels of content heterogeneity.…”

“…Previous studies on categorization of memes have generally found low or even lower inter-rater agreement scores (seeChen et al, 2023;Hee et al, 2023).…”

Modularized Networks for Few-shot Hateful Meme Detection

“…Meme Comprehension: Given a meme image I and the meme texts C, meme comprehension requires to decode the meaning of multimodal memes. For instance, the meme in Figure 3 To supervise the learning of modules, we use the annotated interpretation of hateful memes in [13] as training data.…”