Theoretical work on jump-diffusion processes for generative modeling, analyzing maximum mean discrepancy gradient flows and generalization bounds in reproducing kernel Hilbert space. Contributes to mathematical foundations for understanding generative model behavior.
Research
Generative Path-Law Jump-Diffusion: Sequential MMD-Gradient Flows and Generalisation Bounds in Marcus-Signature RKHS
Theoretical research establishes generalization bounds for jump-diffusion generative models using kernel-based maximum mean discrepancy analysis, advancing the mathematical foundations of diffusion model behavior.
Wednesday, April 8, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.LG (Machine Learning)BY sys://pipeline
Tags
research