New method improves fake news detection with missing modalities

Researchers have developed a new method called Head-wise Modality Decoupling and Regulation (HMDR) to improve fake news detection in multimodal content, particularly when parts of the content are missing. This approach decouples modality representations, regulates their contributions to prevent bias, and promotes invariance learning by aligning representations when one modality is replaced. Experiments on two benchmarks demonstrate that HMDR enhances detection accuracy under missing modalities while preserving performance on complete inputs. AI

IMPACT This research could lead to more robust fake news detection systems capable of handling incomplete multimodal data.

RANK_REASON The cluster contains a research paper detailing a new method for fake news detection. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CL →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• DagsHub
• Gotit.pub
• Hugging Face
• Influence Flower
• Kai Qian
• ScienceCast

• paper
• safety

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