PulseAugur
EN
LIVE 09:20:04

New dataset and model tackle hate speech in Bengali memes

Researchers have introduced Bn-HIB, a new dataset designed to detect hate and inflammatory content within Bengali internet memes. This dataset, containing 3,247 manually annotated memes, is notable for being the first to differentiate between inflammatory content and direct hate speech in Bengali. Alongside the dataset, a Multi-Modal Co-Attention Fusion Model (MCFM) was proposed, which analyzes both visual and textual elements of memes to improve classification accuracy. Experiments indicate that MCFM outperforms existing state-of-the-art models on the Bn-HIB dataset, and the dataset has been made publicly available. AI

IMPACT This research addresses a critical gap in low-resource language NLP, potentially improving content moderation and safety for Bengali-speaking online communities.

RANK_REASON The cluster describes a new academic paper introducing a novel dataset and model for a specific NLP task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

COVERAGE [1]

  1. arXiv cs.CL TIER_1 English(EN) · Rakib Ullah (Sylhet Engineering College), Mominul islam (Daffodil International University), Md Sanjid Hossain (Daffodil International University), Md Ismail Hossain (Daffodil International University) ·

    Detecting Hate and Inflammatory Content in Bengali Memes: A New Multimodal Dataset and Co-Attention Framework

    arXiv:2602.22391v2 Announce Type: replace Abstract: Internet memes have become a dominant form of expression on social media, including within the Bengali speaking community. While often humorous, memes can also be exploited to spread offensive, harmful, and inflammatory content …