Speaker
Description
Anomaly Detection (AD) has recently emerged as a promising approach to the analysis of data collected at the Large Hadron Collider (LHC) in the search for phenomena beyond the Standard Model (SM) of particle physics. AD methods rely on the use of machine learning techniques to learn the characteristics of well-understood Standard Model processes and identify events that exhibit significant deviations from the expected background behavior.
A key advantage of anomaly detection over conventional, model-driven searches at the LHC is its reduced dependence on specific signal hypotheses. Rather than targeting a particular New Physics (NP) process predicted by a given theoretical model, AD aims to identify a broad class of events that are inconsistent with the Standard Model expectation. This model-agnostic character provides a complementary strategy for exploring the vast and largely unconstrained space of possible extensions of the Standard Model.
Such an approach is particularly relevant in the context of several open questions in particle physics, including the nature of dark matter, the origin of the matter–antimatter asymmetry, and the hierarchy problem. By providing sensitivity to unexpected signatures that may not be captured by existing targeted searches, anomaly detection has the potential to broaden the scope of the experimental programme at the LHC.
The work presented in this thesis contributes to this effort by developing and investigating anomaly detection techniques for two distinct applications within the ATLAS experiment. The first concerns the search for New Physics in events characterised by high-transverse-momentum jets and large missing transverse momentum, commonly referred to as the "mono-jet" signature. The second application focuses on the assessment of data quality in the ATLAS Hadronic Tile Calorimeter. These two studies illustrate the versatility of anomaly detection, demonstrating its potential both as a tool for discovering unexpected physics signatures and as a means of identifying anomalous detector behaviour.
| Which topic best fits your talk? | High Energy Physics and Particles |
|---|---|
| Affiliations | Laboratory of Instrumentation and Experimental Particle Physics (PT) |