Classifying anomalies through outer density estimation

Article Article scientifique Article Classifying anomalies through outer density estimation eng 2022-09-06 2022 author Hallin Anna Rutgers U., Piscataway author Isaacson Joshua Fermilab author Kasieczka Gregor Hamburg U. author Krause Claudius Rutgers U., Piscataway author Nachman Benjamin LBNL, Berkeley author Quadfasel Tobias Hamburg U. author Schlaffer Matthias Chicago U. 971999 author Shih David Rutgers U., Piscataway author Sommerhalder Manuel Hamburg U. 055006 10.1103/physrevd.106.055006 2470-0010 2109.00546 https://link.aps.org/doi/10.1103/PhysRevD.106.055006 We propose a new model-agnostic search strategy for physics beyond the standard model (BSM) at the LHC, based on a novel application of neural density estimation to anomaly detection. Our approach, which we call classifying anomalies through outer density estimation (cathode), assumes the BSM signal is localized in a signal region (defined e.g., using invariant mass). By training a conditional density estimator on a collection of additional features outside the signal region, interpolating it into the signal region, and sampling from it, we produce a collection of events that follow the background model. We can then train a classifier to distinguish the data from the events sampled from the background model, thereby approaching the optimal anomaly detector. Using the LHC Olympics R&D dataset, we demonstrate that cathode nearly saturates the best possible performance, and significantly outperforms other approaches that aim to enhance the bump hunt (cwola hunting and anode). Finally, we demonstrate that cathode is very robust against correlations between the features and maintains nearly optimal performance even in this more challenging setting. New physics Neural network Background: model Density Anomaly CERN LHC Coll Performance Estimator Correlation Dewey 500.2 34 Département de physique nucléaire et corpusculaire 132 false Physical review. D 106 5 2ceba287-2603-4d1c-b42e-9739a23eacb9 Article (Published version) Public Article_scientifique_from_unpaywall.pdf 3034403 application/pdf CC BY-4.0 24dc7537de84ba8a84c9e94a12b67459afa0f1d2cd34db0440de7b61c19174e1 openAccess Faculté des sciences / Section de physique / Département de physique nucléaire et corpusculaire

Title	Published in	Access level	OA Policy	Year	Views	Downloads

Classifying anomalies through outer density estimation	Physical review. D			2022	65	147

Archive ouverte UNIGE

Schlaffer, Matthias