MemeScouts@LT-EDI 2026: Asking the Right Questions -- Prompted Weak Supervision for Meme Hate Speech Detection

TL;DR

This paper introduces a prompted weak supervision approach using targeted questions and a quantized vision-language model to improve multilingual meme hate speech detection, outperforming direct classification methods.

Contribution

The paper presents a novel prompted weak supervision method that decomposes meme understanding into targeted questions, enhancing multilingual multimodal hate speech detection performance.

Findings

Outperforms direct VLM classification, especially in Chinese and Hindi.

Achieved 1st place in English, 2nd in Chinese, 3rd in Hindi in the shared task.

Iterative refinement improves generalization and reduces redundancy.

Abstract

Detecting hate speech in memes is challenging due to their multimodal nature and subtle, culturally grounded cues such as sarcasm and context. While recent vision-language models (VLMs) enable joint reasoning over text and images, end-to-end prompting can be brittle, as a single prediction must resolve target, stance, implicitness, and irony. These challenges are amplified in multilingual settings. We propose a prompted weak supervision (PWS) approach that decomposes meme understanding into targeted, question-based labeling functions with constrained answer options for homophobia and transphobia detection in the LT-EDI 2026 shared task. Using a quantized Qwen3-VLM to extract features by answering targeted questions, our method outperforms direct VLM classification, with substantial gains for Chinese and Hindi, ranking 1st in English, 2nd in Chinese, and 3rd in Hindi. Iterative refinement via error-driven LF expansion and feature pruning reduces redundancy and improves generalization. Our results highlight the effectiveness of prompted weak supervision for multilingual multimodal hate speech detection.