Combining outlier analysis algorithms to identify new physics at the LHC

Journal of High Energy Physics

Adam Leinweber¹  Martin White¹  Marco Santoni¹  Paul Jackson¹  Riley Patrick¹  Luc Hendriks²  Sydney Otten³  Sascha Caron⁴  Roberto Ruiz de Austri⁵  Melissa van Beekveld⁶ 

[1] ARC Centre of Excellence for Dark Matter Particle Physics, University of Adelaide, 5005, North Terrace, SA, Australia;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;Gravitation and Astroparticle Physics Amsterdam (GRAPPA), Science Park 904, 1098 XH, Amsterdam, Netherlands;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;Nikhef, Science Park 105, 1098 XG, Amsterdam, Netherlands;Instituto de Física Corpuscular, IFIC-UV/CSIC, Valencia, Spain;Rudolf Peierls Centre for Theoretical Physics, Clarendon Laboratory, 20 Parks Road, OX1 3PU, Oxford, U.K.;High Energy Physics, IMAPP, Radboud University Nijmegen, Heyendaalseweg 135, 6525 AJ, Nijmegen, Netherlands;Nikhef, Science Park 105, 1098 XG, Amsterdam, Netherlands;

关键词: Phenomenological Models;    Supersymmetry Phenomenology;   

DOI  :  10.1007/JHEP09(2021)024

来源: Springer

【 摘 要 】

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 β-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 super- symmetric 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.

【 授权许可】

CC BY