G. Aad, Erlend Aakvaag, Braden Keim Abbott, Sara Abdelhameed, Kira Abeling, Nils Julius Abicht, Haider Abidi, Mohammed Aboelela, Asmaa Aboulhorma, H. Abramowicz, Y. Abulaiti, B. S. Acharya, Anke Ackermann, Claire Adam Bourdarios, Leszek Adamczyk, Sagar Addepalli, Matt Addison, Jahred Adelman, Aytul Adiguzel, Tim Adye, Tony Affolder, Y. Afik, Merve Nazlim Agaras, Anamika Aggarwal, C. Agheorghiesei, F. Ahmadov, S. Ahuja, X. Ai, Giulio Aielli, Arya Aikot, Malak Ait Tamlihat, Brahim Aitbenchikh, Melike Akbiyik, Torsten Akesson, Andrei Akimov, Daiya Akiyama, Nilima Nilesh Akolkar, S. Aktas, Gian Luigi Alberghi, J. Albert, P. Albicocco, Guillaume Lucas Albouy, Sara Alderweireldt, Z. L. Alegria, Martin Aleksa, I. N. Aleksandrov, Calin Alexa, T. Alexopoulos, Fabrizio Alfonsi, Malte Algren, Muhammad Alhroob, Babar Ali, Hanadi Ali, S. Ali, Samuel William Alibocus, M. Aliev, Gianluca Alimonti, Wael Alkakhi, Corentin Allaire, Benedict Allbrooke, Jonny Allen, Julia Frances Allen, Philip Patrick Allport, Alberto Aloisio, F. Alonso, Cristiano Alpigiani, Zainab Mohammad K Alsolami, Adrian Alvarez Fernandez, Mario Alves Cardoso, Mariagrazia Alviggi, M. Aly, Y. Amaral Coutinho, Alessandro Ambler, Christoph Amelung, Maximilian Amerl, Christoph Ames, Dante Amidei, Baktash Amini, Kyle Amirie, Artem Amirkhanov, S. P. Amor Dos Santos, K. R. Amos, Dimitra Amperiadou, Shiwen An, Viktor Ananiev, Christos Anastopoulos, T. Andeen, John Kenneth Anders, Adam Campbell Anderson, A. Andreazza, Stylianos Angelidakis, Aaron Angerami, Alexey Anisenkov, Alberto Annovi, C. Antel, Egor Antipov, Mario Antonelli, Fabio Anulli, Masato Aoki, Takumi Aoki
The ATLAS experiment at the Large Hadron Collider explores the use of modern neural networks for a multi-dimensional calibration of its calorimeter signal defined by clusters of topologically connected cells (topo-clusters). The Bayesian neural network (BNN) approach not only yields a continuous and smooth calibration function that improves performance relative to the standard calibration but also provides uncertainties on the calibrated energies for each topo-cluster. The results obtained by using a trained BNN are compared to the standard local hadronic calibration and to a calibration provided by training a deep neural network. The uncertainties predicted by the BNN are interpreted in the context of a fractional contribution to the systematic uncertainties of the trained calibration. They are also compared to uncertainty predictions obtained from an alternative estimator employing repulsive ensembles.