Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis
Paper β’ 2603.06507 β’ Published β’ 6
Generative approaches for visual synthesis, Invertible deep models for explainable AI, Deep metric and representation learning, self-supervised learning paradigms
RayDer: Scalable Self-Supervised Novel View Synthesis from Real-World Video
Learning Long-term Motion Embeddings for Efficient Kinematics Generation