Showing 1–3 of 3 results for author: Rousselot, A

Generative Invertible Quantum Neural Networks

Authors: Armand Rousselot, Michael Spannowsky

Abstract: Invertible Neural Networks (INN) have become established tools for the simulation and generation of highly complex data. We propose a quantum-gate algorithm for a Quantum Invertible Neural Network (QINN) and apply it to the LHC data of jet-associated production of a Z-boson that decays into leptons, a standard candle process for particle collider precision measurements. We compare the QINN's perfo… ▽ More Invertible Neural Networks (INN) have become established tools for the simulation and generation of highly complex data. We propose a quantum-gate algorithm for a Quantum Invertible Neural Network (QINN) and apply it to the LHC data of jet-associated production of a Z-boson that decays into leptons, a standard candle process for particle collider precision measurements. We compare the QINN's performance for different loss functions and training scenarios. For this task, we find that a hybrid QINN matches the performance of a significantly larger purely classical INN in learning and generating complex data. △ Less

Submitted 22 February, 2024; v1 submitted 24 February, 2023; originally announced February 2023.

Comments: 18 pages, 7 figures Changes in v2: Add references 49-51, provided gitlab link to code repository Changes in v3: Incorporate rebuttal from https://scipost.org/submissions/2302.12906v2/

Report number: IPPP/23/10

Journal ref: SciPost Phys. 16, 146 (2024)

Generative Networks for Precision Enthusiasts

Authors: Anja Butter, Theo Heimel, Sander Hummerich, Tobias Krebs, Tilman Plehn, Armand Rousselot, Sophia Vent

Abstract: Generative networks are opening new avenues in fast event generation for the LHC. We show how generative flow networks can reach percent-level precision for kinematic distributions, how they can be trained jointly with a discriminator, and how this discriminator improves the generation. Our joint training relies on a novel coupling of the two networks which does not require a Nash equilibrium. We… ▽ More Generative networks are opening new avenues in fast event generation for the LHC. We show how generative flow networks can reach percent-level precision for kinematic distributions, how they can be trained jointly with a discriminator, and how this discriminator improves the generation. Our joint training relies on a novel coupling of the two networks which does not require a Nash equilibrium. We then estimate the generation uncertainties through a Bayesian network setup and through conditional data augmentation, while the discriminator ensures that there are no systematic inconsistencies compared to the training data. △ Less

Submitted 19 December, 2022; v1 submitted 22 October, 2021; originally announced October 2021.

Comments: 28 pages, 14 figures

Journal ref: SciPost Phys. 14, 078 (2023)

Invertible Networks or Partons to Detector and Back Again

Authors: Marco Bellagente, Anja Butter, Gregor Kasieczka, Tilman Plehn, Armand Rousselot, Ramon Winterhalder, Lynton Ardizzone, Ullrich Köthe

Abstract: For simulations where the forward and the inverse directions have a physics meaning, invertible neural networks are especially useful. A conditional INN can invert a detector simulation in terms of high-level observables, specifically for ZW production at the LHC. It allows for a per-event statistical interpretation. Next, we allow for a variable number of QCD jets. We unfold detector effects and… ▽ More For simulations where the forward and the inverse directions have a physics meaning, invertible neural networks are especially useful. A conditional INN can invert a detector simulation in terms of high-level observables, specifically for ZW production at the LHC. It allows for a per-event statistical interpretation. Next, we allow for a variable number of QCD jets. We unfold detector effects and QCD radiation to a pre-defined hard process, again with a per-event probabilistic interpretation over parton-level phase space. △ Less

Submitted 1 October, 2020; v1 submitted 11 June, 2020; originally announced June 2020.

Comments: 25 pages, 10 figures

Journal ref: SciPost Phys. 9, 074 (2020)