PC-Droid: Faster diffusion and improved quality for particle cloud generation
This work improves particle cloud generation for high-energy physics simulations, offering faster and higher-quality results, though it builds incrementally on prior diffusion models.
The authors tackled the problem of generating jet particle clouds more efficiently and with higher quality, achieving state-of-the-art performance across all jet types and metrics, with generation times up to two orders of magnitude faster than PC-JeDi and three orders of magnitude faster than Delphes.
Building on the success of PC-JeDi we introduce PC-Droid, a substantially improved diffusion model for the generation of jet particle clouds. By leveraging a new diffusion formulation, studying more recent integration solvers, and training on all jet types simultaneously, we are able to achieve state-of-the-art performance for all types of jets across all evaluation metrics. We study the trade-off between generation speed and quality by comparing two attention based architectures, as well as the potential of consistency distillation to reduce the number of diffusion steps. Both the faster architecture and consistency models demonstrate performance surpassing many competing models, with generation time up to two orders of magnitude faster than PC-JeDi and three orders of magnitude faster than Delphes.