Foundation model for Fast Simulation

This project focuses on leveraging large-scale transformer-based models, inspired by recent advancements in AI such as GPT-3 and DALL-E-2, for fast simulation in high-energy physics (HEP) experiments. Fast simulation plays a critical role in testing hypotheses and generating simulated samples, but the demand for speed and scale necessitates innovative approaches. The student will delve into understanding fast simulation, previous works in this field, and the design of our transformer-based model. Their contribution will involve enhancing data preprocessing and designing specialized loss functions, ultimately aiming to accelerate HEP experiments and advance our understanding of fundamental physics through cutting-edge AI-driven simulations.