multimodalfactneutralPairwise graph formulations are insufficient to capture multi-way cross-modal dependencies in video misinformation detection, whereas hypergraphs offer a suitable representationArtificial Intelligence02 Aug 2026http://arxiv.org/abs/2607.28375v1