M1: Optimization of Event Selection for the Fully Leptonic WZ Vector Boson Scattering Analysisin ATLAS Using Artificial Intelligence Techniques

Scientific Context :
The Standard Model (SM) of particle physics provides a remarkably precise description of fundamental
interactions among elementary particles. Among the processes particularly sensitive to possible new
physics are Vector Boson Scattering (VBS) events, which probe the structure of electroweak symmetry breaking and multi-boson interactions.
The ATLAS experiment at the Large Hadron Collider (LHC) at CERN measures these processes at unprecedented energies and in specific final states. In particular, the fully leptonic WZ channel offers
a clean but rare signature, requiring an efficient event selection strategy to isolate the signal from the
dominant backgrounds.

Internship Objective :
The goal of this internship is to explore and optimize the event selection strategy used in the fully leptonic WZ VBS analysis.
The work will involve:
• Studying the impact of various selection cuts on key kinematic observables (invariant masses,
angular variables, rapidities, etc.);
• Investigating potential improvements guided by artificial intelligence (AI) methods, such as variable ranking algorithms, lightweight neural networks, or automated cut optimization;
• Quantifying the expected gain in analysis sensitivity based on the proposed optimizations.

Methodology :
The work will be carried out within the ATLAS group at LAPP (Annecy, France), in close collaboration with physicists involved in multi-boson and VBS analyses.
The main steps of the internship will include:
1. Familiarization with the ATLAS analysis framework and software tools (ROOT, Python,
local analysis framework, and handling of simulated ntuple datasets);
2. Study of discriminating observables between signal and main backgrounds (non-VBS WZ,
top, ZZ, etc.);
3. Exploration of optimization strategies for event selection, using both classical approaches
(variable scans) and AI-assisted methods (feature importance, decision trees, automated
optimization);
4. Evaluation of performance improvements in terms of signal-to-background ratio and
statistical significance;
5. Preparation of a final report and presentation to the ATLAS group at LAPP.

Supervision and Work Environment :
The internship will take place within the ATLAS group at LAPP, composed of around ten senior
physicists, several post-doctoral researchers, and PhD students.
Direct supervision will be provided by:
• Three senior physicists involved in electroweak and multi-boson analyses,
• One post-doctoral researcher and one PhD student specialized in the fully leptonic WZ
analysis.
The student will work in a dynamic, international research environment, with opportunities for interaction
with the CERN community (located 50 km from Annecy).

Required/desired skills :
• Basic knowledge of particle physics and the Standard Model;
• Interest in experimental data analysis;
• Familiarity with scientific programming (Python, C++, ROOT);
• Interest in artificial intelligence techniques applied to physics.

Duration and location:
• Duration: ~12 weeks
• Location: Laboratoire d’Annecy de Physique des Particules (LAPP), Annecy, France
• Supervising group: ATLAS group at LAPP (O. Arnaez, A. Carnelli, I. Koletsou, E. Sauvan).

Informations:

Expérience: ATLAS

Responsable de stage: Olivier ARNAEZ

Email: olivier.arnaez@lapp.in2p3.fr