Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

T: Fachverband Teilchenphysik

T 54: Data, AI, Computing, Electronics V (Anomaly Detection, Event Selection)

T 54.7: Vortrag

Mittwoch, 2. April 2025, 17:45–18:00, VG 2.101

Enhancing the identification of HH → bbbb by Triplet Learning — •Bao Tai Le, Lars Linden, Otmar Biebel, Stephanie Götz, Celine Stauch, Valerio D’Amico, and Tim Rexrodt — Ludwig-Maximilians-Universität, München, Deutschland

In recent years various machine learning techniques have proven to be quite successful in particle physics replacing old methodology and introducing new ways of thinking. One of those ways is Triplet Training. Its appeal comes from its resilience against noisy data by forming a more salient feature space leading to better categorization performances across many different categorization architectures.The production of a pair of Higgs bosons is possible due to the Higgs self interaction. However, the cross section of this process is tiny and the largest branching ratio of the Higgs decay involves bottom quarks which are also abundantly produced by strong interaction in proton-proton collisions. Even though bottom quark jets can be identified e.g. by secondary decay vertices, it is an experimental challenge to maintain a high efficiency to identify the four b-quark jets from a HH→4b event. Due to the resilience of Triplet Learning against noisy data its application seems promising for enhancing the identification efficiency of HH → 4b events.

Keywords: Higgs Particle; Machine Learning; Particle Physics; Data Analysis; Standard Model