BEGIN:VCALENDAR VERSION:2.0 PRODID:-//linuxsoftware.nz//NONSGML Joyous v1.4//EN BEGIN:VEVENT SUMMARY:New Physics searches at the LHC using Anomaly Detection DTSTART:20241128T160000Z DTEND:20241128T180000Z DTSTAMP:20260725T173044Z UID:d908cde6-53eb-4c2d-bc82-d5e2b408f38a SEQUENCE:2 CREATED:20241122T221022Z DESCRIPTION:Collider experiments provide a powerful means of exploring hig h-energy physics and identifying potential signatures of new physics. This study contributes to these efforts in a model-agnostic way\, using semi-s upervised learning approaches for anomaly detection. Various searches for new physics focus on fully hadronic final states\, with jets serving as pr obes for potential signals. This work specifically targets anomaly detecti on at the jet level. Each jet is represented as a graph\, with nodes corre sponding to its hadronic constituents. Simulated datasets of Dark Jets eve nts\, framed within a dark matter model\, serve as the benchmark signal\, where a heavy vector boson Z'\; mediator connects a Standard Model qua rk pair with a pair of dark quarks. These quarks then shower and hadronize \, producing dark jets. The background consists of QCD dijet events. The o bjective is to extract a vector embedding that maps high-dimensional graph information into a low-dimensional vector using convolution and pooling m echanisms. This embedding serves as input to an AD method\, such as DeepSV DDs and Autoencoders\, allowing for jet prediction and classification base d on anomaly scores. Performance comparisons are conducted against baselin e deep learning approaches. LAST-MODIFIED:20241127T114325Z LOCATION:Online URL:http://df.vps.tecnico.ulisboa.pt/en/events/new-physics-searches-at-the -lhc-using-anomaly-detection/ X-ALT-DESC;FMTTYPE=text/html:

Collider experiment s provide a powerful means of exploring high-energy physics and identifyin g potential signatures of new physics. This study contributes to these eff orts in a model-agnostic way\, using semi-supervised learning approaches f or anomaly detection. Various searches for new physics focus on fully hadr onic final states\, with jets serving as probes for potential signals.

This work specifically targets anomaly detection at the jet level. Each jet is represented as a graph\, with nodes corresponding to its hadro nic constituents. Simulated datasets of Dark Jets events\, framed within a dark matter model\, serve as the benchmark signal\, where a heavy vector boson Z'\; mediator connects a Standard Model quark pair with a pair o f dark quarks.

These quarks then shower and hadronize\, producin g dark jets. The background consists of QCD dijet events. The objective is to extract a vector embedding that maps high-dimensional graph informatio n into a low-dimensional vector using convolution and pooling mechanisms. This embedding serves as input to an AD method\, such as DeepSVDDs and Aut oencoders\, allowing for jet prediction and classification based on anomal y scores. Performance comparisons are conducted against baseline deep lear ning approaches.

END:VEVENT END:VCALENDAR