NOUMBISSI Ange Landry

TopoFusion-TR Hyper is an innovative multimodal AI framework that combines deep learning, differential geometry, and persistent homology to learn robust and structurally consistent representations across different modalities, including images, text, audio, and video. By moving beyond traditional Euclidean similarity-based alignment, it preserves the intrinsic geometric and topological structure of data while improving cross-modal semantic alignment and resistance to noise. Integrated with Meta's Muse Spark ecosystem, TopoFusion aims to advance reliable multimodal reasoning for applications in document understanding, healthcare, robotics, and physical AI.