Calo-VQ: Vector-Quantized Two-Stage Generative Model in Calorimeter Simulation
This provides accelerated simulation for particle physics experiments, though it appears to be an incremental adaptation of existing VQ-VAE techniques to a specific domain.
The paper tackles fast simulation of calorimeter detector response by developing a two-stage generative model using vector-quantized variational autoencoders, achieving a 2000x speed improvement over conventional methods with generation times in milliseconds.
We introduce a novel machine learning method developed for the fast simulation of calorimeter detector response, adapting vector-quantized variational autoencoder (VQ-VAE). Our model adopts a two-stage generation strategy: initially compressing geometry-aware calorimeter data into a discrete latent space, followed by the application of a sequence model to learn and generate the latent tokens. Extensive experimentation on the Calo-challenge dataset underscores the efficiency of our approach, showcasing a remarkable improvement in the generation speed compared with conventional method by a factor of 2000. Remarkably, our model achieves the generation of calorimeter showers within milliseconds. Furthermore, comprehensive quantitative evaluations across various metrics are performed to validate physics performance of generation.