Meanmemes (left), benignimageconfounders (middle) andbenigntextconfounders (right).

This work proposes a new challenge set for multimodal classification, focusing on detecting hate speech in multimodal memes. It is constructed such that unimodal models struggle and only multimodal models can succeed: difficult examples ("benign confounders") are added to the dataset to make it hard to rely on unimodal signals. The task requires subtle reasoning, yet is straightforward to evaluate as a binary classification problem. We provide baseline performance numbers for unimodal models, as well as for multimodal models with various degrees of sophistication. We find that state-of-the-art methods perform poorly compared to humans, illustrating the difficulty of the task and highlighting the challenge that this important problem poses to the community.

Memes are widely used for humor and cultural commentary, but they are increasingly exploited to spread hateful content. Due to their multimodal nature, hateful memes often evade traditional text-only or image-only detection systems, particularly when they employ subtle or coded references. To address these challenges, we propose a multimodal hate detection framework that integrates key components: OCR to extract embedded text, captioning to describe visual content neutrally, sub-label classification for granular categorization of hateful content, RAG for contextually relevant retrieval, and VQA for iterative analysis of symbolic and contextual cues. This enables the framework to uncover latent signals that simpler pipelines fail to detect. Experimental results on the Facebook Hateful Memes dataset reveal that the proposed framework exceeds the performance of unimodal and conventional mul-timodal models in both accuracy and AUC-ROC.

Different people may have different opinions for the same memes. Must the opposite of the hateful speech be harmless ones? Hate speech detection is NOT a strictly binary classification problem. Hence, the modeling for the task in the paper is inaccurate. Could the authors present a candidate solution for it?

This work proposes a new challenge set for multimodal classification, focusing on detecting hate speech in multimodal memes. It is constructed such that unimodal models struggle and only multimodal models can succeed: difficult examples ("benign confounders") are added to the dataset to make it hard to rely on unimodal signals. The task requires subtle reasoning, yet is straightforward to evaluate as a binary classification problem. We provide baseline performance numbers for unimodal models, as well as for multimodal models with various degrees of sophistication. We find that state-of-the-art methods perform poorly compared to humans, illustrating the difficulty of the task and highlighting the challenge that this important problem poses to the community.

Moderation of social media content is currently a highly manual task, yet there is too much content posted daily to do so effectively. With the advent of a number of multimodal models, there is the potential to reduce the amount of manual labor for this task. In this work, we aim to explore different models and determine what is most effective for the Hateful Memes Challenge, a challenge by Meta designed to further machine learning research in content moderation. Specifically, we explore the differences between early fusion and late fusion models in classifying multimodal memes containing text and images. We first implement a baseline using unimodal models for text and images separately using BERT and ResNet-152, respectively. The outputs from these unimodal models were then concatenated together to create a late fusion model. In terms of early fusion models, we implement ConcatBERT, VisualBERT, ViLT, CLIP, and BridgeTower. It was found that late fusion performed significantly worse than early fusion models, with the best performing model being CLIP which achieved an AUROC of 70.06. The code for this work is available at https://github.com/bzhao18/CS-7643-Project.

State-of-the-art image and text classification models, such as Convolutional Neural Networks and Transformers, have long been able to classify their respective unimodal reasoning satisfactorily with accuracy close to or exceeding human accuracy. However, images embedded with text, such as hateful memes, are hard to classify using unimodal reasoning when difficult examples, such as benign confounders, are incorporated into the data set. We attempt to generate more labeled memes in addition to the Hateful Memes data set from Facebook AI, based on the framework of a winning team from the Hateful Meme Challenge. To increase the number of labeled memes, we explore semi-supervised learning using pseudo-labels for newly introduced, unlabeled memes gathered from the Memotion Dataset 7K. We find that the semi-supervised learning task on unlabeled data required human intervention and filtering and that adding a limited amount of new data yields no extra classification performance.

Hateful Meme Challenge proposed by Facebook AI has attracted contestants around the world. The challenge focuses on detecting hateful speech in multimodal memes. Various state-of-the-art deep learning models have been applied to this problem and the performance on challenge's Figure 1: Examples of Hateful Memes [3] leaderboard has also been constantly improved. In this paper, we enhance the hateful detection framework, including utilizing Detectron for feature extraction, exploring different the gap between the best model and human is still large. A setups of VisualBERT and UNITER models with different recent research comparing models of hateful speech detection loss functions, researching the association between the in multimodal memes and human shows an accuracy hateful memes and the sensitive text features, and finally of 64.73% and 84.7%[12], leaving much room for further building ensemble method to boost model performance.

The recently introduced hateful meme challenge demonstrates the difficulty of determining whether a meme is hateful or not. Specifically, both unimodal language models and multimodal vision-language models cannot reach the human level of performance. Motivated by the need to model the contrast between the image content and the overlayed text, we suggest applying an off-the-shelf image captioning tool in order to capture the first. We demonstrate that the incorporation of such automatic captions during fine-tuning improves the results for various unimodal and multimodal models. Moreover, in the unimodal case, continuing the pre-training of language models on augmented and original caption pairs, is highly beneficial to the classification accuracy.

Memes on the Internet are often harmless and sometimes amusing. However, by using certain types of images, text, or combinations of both, the seemingly harmless meme becomes a multimodal type of hate speech -- a hateful meme. The Hateful Memes Challenge is a first-of-its-kind competition which focuses on detecting hate speech in multimodal memes and it proposes a new data set containing 10,000+ new examples of multimodal content. We utilize VisualBERT -- which meant to be the BERT of vision and language -- that was trained multimodally on images and captions and apply Ensemble Learning. Our approach achieves 0.811 AUROC with an accuracy of 0.765 on the challenge test set and placed third out of 3,173 participants in the Hateful Memes Challenge.