Beyond Binary Classification: Detecting Fine-Grained Sexism in Social Media Videos

2026-02-17

AI Summary

This study addresses the limitations of existing automated detection tools, which are predominantly confined to binary classification and struggle to capture the nuanced and diverse manifestations of gender-based discrimination in social media videos. To advance this field, the authors introduce FineMuSe, the first multimodal Spanish-language dataset supporting fine-grained annotation, along with a hierarchical taxonomy that encompasses distinct types of gender discrimination, non-discriminatory content, and ironic or humorous expressions. The paper evaluates the performance of multimodal large language models on both binary and fine-grained classification tasks. Experimental results indicate that these models approach human-level accuracy in identifying subtle forms of gender discrimination, yet they still face challenges when confronted with co-occurring visual and textual cues signaling multiple overlapping discriminatory types. This work thus shifts the paradigm from coarse-grained detection toward context-sensitive, fine-grained analysis of gender bias.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Online sexism appears in various forms, which makes its detection challenging. Although automated tools can enhance the identification of sexist content, they are often restricted to binary classification. Consequently, more subtle manifestations of sexism may remain undetected due to the lack of fine-grained, context-sensitive labels. To address this issue, we make the following contributions: (1) we present FineMuSe, a new multimodal sexism detection dataset in Spanish that includes both binary and fine-grained annotations; (2) we introduce a comprehensive hierarchical taxonomy that encompasses forms of sexism, non-sexism, and rhetorical devices of irony and humor; and (3) we evaluate a wide range of LLMs for both binary and fine-grained sexism detection. Our findings indicate that multimodal LLMs perform competitively with human annotators in identifying nuanced forms of sexism; however, they struggle to capture co-occurring sexist types when these are conveyed through visual cues.

Problem

• sexism detection
• fine-grained classification
• multimodal social media
• context-sensitive labeling
• online gender bias

Innovation

• fine-grained sexism detection
• multimodal dataset
• hierarchical taxonomy
• large language models
• social media video