Balazs, C. et al, Mamuzic, J., & Ruiz de Austri, R. (2021). A comparison of optimisation algorithms for high-dimensional particle and astrophysics applications. J. High Energy Phys., 05(5), 108–46pp.
Abstract: Optimisation problems are ubiquitous in particle and astrophysics, and involve locating the optimum of a complicated function of many parameters that may be computationally expensive to evaluate. We describe a number of global optimisation algorithms that are not yet widely used in particle astrophysics, benchmark them against random sampling and existing techniques, and perform a detailed comparison of their performance on a range of test functions. These include four analytic test functions of varying dimensionality, and a realistic example derived from a recent global fit of weak-scale supersymmetry. Although the best algorithm to use depends on the function being investigated, we are able to present general conclusions about the relative merits of random sampling, Differential Evolution, Particle Swarm Optimisation, the Covariance Matrix Adaptation Evolution Strategy, Bayesian Optimisation, Grey Wolf Optimisation, and the PyGMO Artificial Bee Colony, Gaussian Particle Filter and Adaptive Memory Programming for Global Optimisation algorithms.
|
Hirsch, M., Krauss, M. E., Opferkuch, T., Porod, W., & Staub, F. (2016). A constrained supersymmetric left-right model. J. High Energy Phys., 03(3), 009–22pp.
Abstract: We present a supersymmetric left-right model which predicts gauge coupling unification close to the string scale and extra vector bosons at the TeV scale. The subtleties in constructing a model which is in agreement with the measured quark masses and mixing for such a low left-right breaking scale are discussed. It is shown that in the constrained version of this model radiative breaking of the gauge symmetries is possible and a SM-like Higgs is obtained. Additional CP-even scalars of a similar mass or even much lighter are possible. The expected mass hierarchies for the supersymmetric states differ clearly from those of the constrained MSSM. In particular, the lightest down-type squark, which is a mixture of the sbottom and extra vector-like states, is always lighter than the stop. We also comment on the model's capability to explain current anomalies observed at the LHC.
|
Chala, M., Delgado, A., Nardini, G., & Quiros, M. (2017). A light sneutrino rescues the light stop. J. High Energy Phys., 04(4), 097–22pp.
Abstract: Stop searches in supersymmetric frameworks with R-parity conservation usually assume the lightest neutralino to be the lightest supersymmetric particle. In this paper we consider an alternative scenario in which the left-handed tau sneutrino is lighter than neutralinos and stable at collider scales, but possibly unstable at cosmological scales. Moreover the (mostly right-handed) stop (t) over tilde is lighter than all electroweakinos, and heavier than the scalars of the third generation lepton doublet, whose charged component, (T) over tilde, is heavier than the neutral one, (v) over tilde. The remaining supersymmetric particles are decoupled from the stop phenomenology. In most of the parameter space, the relevant stop decays are only into t (T) over tildeT, t (v) over tildev and b (v) over tildeT via off-shell electroweakinos. We constrain the branching ratios of these decays by recasting the most sensitive stop searches. Due to the “double invisible” kinematics of the (t) over tilde -> t (v) over tildev process, and the low efficiency in tagging the t (T) over tildeT decay products, light stops are generically allowed. In the minimal supersymmetric standard model with similar to 100 GeV sneutrinos, stops with masses as small as similar to 350 GeV turn out to be allowed at 95% CL.
|
Feroz, F., Cranmer, K., Hobson, M., Ruiz de Austri, R., & Trotta, R. (2011). Challenges of profile likelihood evaluation in multi-dimensional SUSY scans. J. High Energy Phys., 06(6), 042–23pp.
Abstract: Statistical inference of the fundamental parameters of supersymmetric theories is a challenging and active endeavor. Several sophisticated algorithms have been employed to this end. While Markov-Chain Monte Carlo (MCMC) and nested sampling techniques are geared towards Bayesian inference, they have also been used to estimate frequentist confidence intervals based on the profile likelihood ratio. We investigate the performance and appropriate configuration of MULTINEST, a nested sampling based algorithm, when used for profile likelihood-based analyses both on toy models and on the parameter space of the Constrained MSSM. We find that while the standard configuration previously used in the literarture is appropriate for an accurate reconstruction of the Bayesian posterior, the profile likelihood is poorly approximated. We identify a more appropriate MULTINEST configuration for profile likelihood analyses, which gives an excellent exploration of the profile likelihood (albeit at a larger computational cost), including the identification of the global maximum likelihood value. We conclude that with the appropriate configuration MULTINEST is a suitable tool for profile likelihood studies, indicating previous claims to the contrary are not well founded.
|
van Beekveld, M., Caron, S., Hendriks, L., Jackson, P., Leinweber, A., Otten, S., et al. (2021). Combining outlier analysis algorithms to identify new physics at the LHC. J. High Energy Phys., 09(9), 024–33pp.
Abstract: The lack of evidence for new physics at the Large Hadron Collider so far has prompted the development of model-independent search techniques. In this study, we compare the anomaly scores of a variety of anomaly detection techniques: an isolation forest, a Gaussian mixture model, a static autoencoder, and a beta-variational autoencoder (VAE), where we define the reconstruction loss of the latter as a weighted combination of regression and classification terms. We apply these algorithms to the 4-vectors of simulated LHC data, but also investigate the performance when the non-VAE algorithms are applied to the latent space variables created by the VAE. In addition, we assess the performance when the anomaly scores of these algorithms are combined in various ways. Using supersymmetric benchmark points, we find that the logical AND combination of the anomaly scores yielded from algorithms trained in the latent space of the VAE is the most effective discriminator of all methods tested.
|