# More Than Sum of Its Parts: Deciphering Intent Shifts in Multimodal Hate Speech Detection

arXiv: https://arxiv.org/abs/2603.21298 · HTML: https://mlsanity.com/paper/2603.21298/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 57 references that are in the map.
- First posted on arXiv 22 Mar 2026.

## Builds on (most cited first, 22 of 22)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [Align before Fuse: Vision and Language Representation Learning with Momentum Distillation](https://mlsanity.com/paper/2107.07651/) (2021)
- [Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debate](https://mlsanity.com/paper/2305.19118/) (2023)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [The Hateful Memes Challenge: Detecting Hate Speech in Multimodal Memes](https://mlsanity.com/paper/2005.04790/) (2020)
- [OBELICS: An Open Web-Scale Filtered Dataset of Interleaved Image-Text Documents](https://mlsanity.com/paper/2306.16527/) (2023)
- [Supervised Multimodal Bitransformers for Classifying Images and Text](https://mlsanity.com/paper/1909.02950/) (2019)
- [Exploring Hate Speech Detection in Multimodal Publications](https://mlsanity.com/paper/1910.03814/) (2019)
- [Detecting Harmful Memes and Their Targets](https://mlsanity.com/paper/2110.00413/) (2021)
- [Prompting for Multimodal Hateful Meme Classification](https://mlsanity.com/paper/2302.04156/) (2023)
- [Pro-Cap: Leveraging a Frozen Vision-Language Model for Hateful Meme Detection](https://mlsanity.com/paper/2308.08088/) (2023)
- [Skyra: AI-Generated Video Detection via Grounded Artifact Reasoning](https://mlsanity.com/paper/2512.15693/) (2025)
- [Improving Hateful Meme Detection through Retrieval-Guided Contrastive Learning](https://mlsanity.com/paper/2311.08110/) (2023)
- [CoSyn: Detecting Implicit Hate Speech in Online Conversations Using a Context Synergized Hyperbolic Network](https://mlsanity.com/paper/2303.03387/) (2023)
- [Debate-to-Detect: Reformulating Misinformation Detection as a Real-World Debate with Large Language Models](https://mlsanity.com/paper/2505.18596/) (2025)
- [Target Span Detection for Implicit Harmful Content](https://mlsanity.com/paper/2403.19836/) (2024)
- [Learning Counterfactually Decoupled Attention for Open-World Model Attribution](https://mlsanity.com/paper/2506.23074/) (2025)
- [Cracking the Code: Enhancing Implicit Hate Speech Detection through Coding Classification](https://mlsanity.com/paper/2506.04693/) (2025)

## Abstract (arXiv)

Combating hate speech on social media is critical for securing cyberspace, yet relies heavily on the efficacy of automated detection systems. As content formats evolve, hate speech is transitioning from solely plain text to complex multimodal expressions, making implicit attacks harder to spot. Current systems, however, often falter on these subtle cases, as they struggle with multimodal content where the emergent meaning transcends the aggregation of individual modalities. To bridge this gap, we move beyond binary classification to characterize semantic intent shifts where modalities interact to construct implicit hate from benign cues or neutralize toxicity through semantic inversion. Guided by this fine-grained formulation, we curate the Hate via Vision-Language Interplay (H-VLI) benchmark where the true intent hinges on the intricate interplay of modalities rather than overt visual or textual slurs. To effectively decipher these complex cues, we further propose the Asymmetric Reasoning via Courtroom Agent DEbate (ARCADE) framework. By simulating a judicial process where agents actively argue for accusation and defense, ARCADE forces the model to scrutinize deep semantic cues before reaching a verdict. Extensive experiments demonstrate that ARCADE significantly outperforms state-of-the-art baselines on H-VLI, particularly for challenging implicit cases, while maintaining competitive performance on established benchmarks. Our code and data are available at: https://github.com/Sayur1n/H-VLI