Speaker
Dr
Vidya Manian
(University of Puerto Rico, Mayaguez)
Description
Score based and consistency diffusion models are presented for generating jet images, focusing on high-fidelity synthesis for high energy Physics applications. Using the JetNet dataset, the diffusion models are trained to learn the visual representation of jet kinematics. The results demonstrate that consistency models achieve significantly lower Fréchet inception distance measures compared to score-based models, indicating improved image quality and generation stability. Unlike methods based on latent distributions, this approach operates directly in image space. Furthermore, the efficacy of jet image generation is demonstrated using jet tagging and other metrics to highlight the strengths of image-based jet generative modeling.
Authors
Mr
Victor Martinez
(University of Puerto Rico, Mayaguez)
Dr
Vidya Manian
(University of Puerto Rico, Mayaguez)
Co-author
Dr
Sudhir Malik
(University of Puerto Rico, Mayaguez)
