Skip to main content

Research timeline

Related research and updates

Public articles linked to the same research event.

arXiv

GraphBind binds multimodal semantics through graph topology, reaching up to 28.1% relative gains over 11 baselines

The work proposes GraphBind, a topology-driven approach that binds rich modality information into a unified shared space for multimodal graph foundation models, using topology to organize self semantics and reliable neighborhood semantics and adapting that space to discriminative and generative tasks through lightweight interfaces; experiments against 11 representative baselines report leading performance on both task types, with relative improvements of up to 28.1% over the strongest baseline.