May 25 – 29, 2026
Chulalongkorn University
Asia/Bangkok timezone

Training and Deployment of Transformers in ATLAS

May 28, 2026, 3:15 PM
1h
Sala Phra Kieo, Second Floor

Sala Phra Kieo, Second Floor

Board: 308-Th
Poster Presentation Track 3 - Offline data processing Poster

Speaker

Alessandro Zaio (INFN e Universita Genova (IT))

Description

Transformer architectures have rapidly become the state-of-the-art approach for machine-learning models across many domains in science, offering unprecedented performance on complex, high-dimensional tasks. Their adoption within the ATLAS experiment, starting with their usage for flavour tagging, has opened new opportunities, but also introduced substantial challenges regarding large-scale training, infrastructure integration, and deployment within established software and production system of ATLAS. In this talk, we will provide an overview on how we train some of the transformer-based models in ATLAS, including the preparation of the dataset as well as hardware-specific constraints. We will also outline the latest developments in deployment workflows and their integration in the ATLAS offline production system.

Authors

Alessandro Zaio (INFN e Universita Genova (IT)) Alexander Froch (Universite de Geneve (CH)) Angela Maria Burger (Centre National de la Recherche Scientifique (FR)) Diego Baron (The University of Manchester (GB)) Romain Bouquet (INFN - University of Genova (Italy)) Valentina 🐧 Vecchio (University of Manchester)

Presentation materials

There are no materials yet.