Publications

Order by: publication date number of citations

• Phase space sampling and inference from weighted events with autoregressive flows

  Bob Stienen, Rob Verheyen

  SciPost Phys. 10, 038 (2021) · published 17 February 2021 | Toggle abstract · pdf

  We explore the use of autoregressive flows, a type of generative model with tractable likelihood, as a means of efficient generation of physical particle collider events. The usual maximum likelihood loss function is supplemented by an event weight, allowing for inference from event samples with variable, and even negative event weights. To illustrate the efficacy of the model, we perform experiments with leading-order top pair production events at an electron collider with importance sampling weights, and with next-to-leading-order top pair production events at the LHC that involve negative weights.

  35 citations

• Deep-learning jets with uncertainties and more

  Sven Bollweg, Manuel Haussmann, Gregor Kasieczka, Michel Luchmann, Tilman Plehn, Jennifer Thompson

  SciPost Phys. 8, 006 (2020) · published 16 January 2020 | Toggle abstract · pdf

  Bayesian neural networks allow us to keep track of uncertainties, for example in top tagging, by learning a tagger output together with an error band. We illustrate the main features of Bayesian versions of established deep-learning taggers. We show how they capture statistical uncertainties from finite training samples, systematics related to the jet energy scale, and stability issues through pile-up. Altogether, Bayesian networks offer many new handles to understand and control deep learning at the LHC without introducing a visible prior effect and without compromising the network performance.

  34 citations

• STrEAMlining EFT Matching

  Timothy Cohen, Xiaochuan Lu, Zhengkang Zhang

  SciPost Phys. 10, 098 (2021) · published 3 May 2021 | Toggle abstract · pdf

  This paper presents STrEAM (SuperTrace Evaluation Automated for Matching), a Mathematica package that calculates all functional supertraces which arise when matching a generic UV model onto a relativistic Effective Field Theory (EFT) at one loop and to arbitrary order in the heavy mass expansion. STrEAM implements the covariant derivative expansion to automate the most tedious step of the streamlined functional matching prescription presented in arXiv:2011.02484 . The code and an example notebook are available at https://www.github.com/EFTMatching/STrEAM .

  33 citations

• Dark matter, fine-tuning and $(g-2)_{\mu}$ in the pMSSM

  Melissa van Beekveld, Wim Beenakker, Marrit Schutten, Jeremy De Wit

  SciPost Phys. 11, 049 (2021) · published 7 September 2021 | Toggle abstract · pdf

  In this paper we perform for the first time an in-depth analysis of the spectra in the phenomenological supersymmetric Standard Model that simultaneously offer an explanation for the $(g-2)_{\mu}$ discrepancy $\Delta a_{\mu}$, result in the right dark-matter relic density $\Omega_{\rm DM} h^2$ and are minimally fine-tuned. The resulting spectra may be obtained from [1]. To discuss the experimental exclusion potential for our models, we analyse the resulting LHC phenomenology as well as the sensitivity of dark-matter direct detection experiments to these spectra. We find that the latter type of experiments with sensitivity to the spin-dependent dark-matter\,--\,nucleon scattering cross section $\sigma_{\rm SD,p}$ will probe all of our found solutions.

  33 citations

• Towards the renormalisation of the Standard Model effective field theory to dimension eight: Bosonic interactions I

  Mikael Chala, Guilherme Guedes, Maria Ramos, Jose Santiago

  SciPost Phys. 11, 065 (2021) · published 21 September 2021 | Toggle abstract · pdf

  We compute the one-loop renormalisation group running of the bosonic Standard Model effective operators to order $v^4/\Lambda^4$, with $v\sim 246$ GeV being the electroweak scale and $\Lambda$ the unknown new physics threshold. We concentrate on the effects triggered by pairs of the leading dimension-six interactions, namely those that can arise at tree level in weakly-coupled ultraviolet completions of the Standard Model. We highlight some interesting consequences, including the interplay between positivity bounds and the form of the anomalous dimensions; the non renormalisation of the $S$ and $U$ parameters; or the importance of radiative corrections to the Higgs potential for the electroweak phase transition. As a byproduct of this work, we provide a complete Green basis of operators involving only the Higgs and derivatives at dimension-eight, comprising 13 redundant interactions.

  33 citations

• QED at NNLO with McMule

  P. Banerjee, T. Engel, A. Signer, Y. Ulrich

  SciPost Phys. 9, 027 (2020) · published 31 August 2020 | Toggle abstract · pdf

  McMule is a framework for fully differential higher-order QED calculations of scattering and decay processes involving leptons. It keeps finite lepton masses, which regularises collinear singularities. Soft singularities are treated with dimensional regularisation and using FKS$^\ell$ subtraction. We describe the implementation of the framework in Fortran 95, list the processes that are currently implemented, and give instructions on how to run the code. In addition, we present new phenomenological results for muon-electron scattering and lepton-proton scattering, including the dominant NNLO corrections. While the applications presented focus on MUonE, MUSE, and P2, the code can be used for a large number of planned and running experiments.

  33 citations

• Neutrino experiments probe hadrophilic light dark matter

  Yohei Ema, Filippo Sala, Ryosuke Sato

  SciPost Phys. 10, 072 (2021) · published 24 March 2021 | Toggle abstract · pdf

  We use Super-K data to place new strong limits on interactions of sub-GeV Dark Matter (DM) with nuclei, that rely on the DM flux inevitably induced by cosmic-ray upscatterings. We derive analogous sensitivities at Hyper-K and DUNE and compare them with others, e.g. at JUNO. Using simplified models, we find that our proposal tests genuinely new parameter space, allowed both by theoretical consistency and by other direct detection experiments, cosmology, meson decays and our recast of monojet. Our results thus motivate and shape a new physics case for any large volume detector sensitive to nuclear recoils.

  32 citations

• Better latent spaces for better autoencoders

  Barry M. Dillon, Tilman Plehn, Christof Sauer, Peter Sorrenson

  SciPost Phys. 11, 061 (2021) · published 17 September 2021 | Toggle abstract · pdf

  Autoencoders as tools behind anomaly searches at the LHC have the structural problem that they only work in one direction, extracting jets with higher complexity but not the other way around. To address this, we derive classifiers from the latent space of (variational) autoencoders, specifically in Gaussian mixture and Dirichlet latent spaces. In particular, the Dirichlet setup solves the problem and improves both the performance and the interpretability of the networks.

  32 citations

• A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty

  Benjamin Nachman

  SciPost Phys. 8, 090 (2020) · published 23 June 2020 | Toggle abstract · pdf

  Deep learning tools can incorporate all of the available information into a search for new particles, thus making the best use of the available data. This paper reviews how to optimally integrate information with deep learning and explicitly describes the corresponding sources of uncertainty. Simple illustrative examples show how these concepts can be applied in practice.

  30 citations

• Mass agnostic jet taggers

  Layne Bradshaw, Rashmish K. Mishra, Andrea Mitridate, Bryan Ostdiek

  SciPost Phys. 8, 011 (2020) · published 24 January 2020 | Toggle abstract · pdf

  Searching for new physics in large data sets needs a balance between two competing effects---signal identification vs background distortion. In this work, we perform a systematic study of both single variable and multivariate jet tagging methods that aim for this balance. The methods preserve the shape of the background distribution by either augmenting the training procedure or the data itself. Multiple quantitative metrics to compare the methods are considered, for tagging 2-, 3-, or 4-prong jets from the QCD background. This is the first study to show that the data augmentation techniques of Planing and PCA based scaling deliver similar performance as the augmented training techniques of Adversarial NN and uBoost, but are both easier to implement and computationally cheaper.

  30 citations