Transformer-based models to fast-simulate the LHCb ECAL detector

Transformer

The Transformer architecture is freely inspired by Vaswani et al. [arXiv:1706.03762] and Dosovitskiy et al. [arXiv:2010.11929].

Discriminator

The Discriminator is implemented through the Deep Sets model proposed by Zaheer et al. [arXiv:1703.06114] and its architecture is freely inspired by what developed by the ATLAS Collaboration for flavor tagging [ATL-PHYS-PUB-2020-014].

Credits

Transformer implementation freely inspired by the TensorFlow tutorial Neural machine translation with a Transformer and Keras.