Supersymmetry

Pioneering the search for the supersymmetric partner of the top quark (and the charm quark) in LHC’s Run 1

Machine Learning

Conventional algorithms to reconstruct the tracks of charged particles are computationally intractable in the planned High-Luminosity LHC running. We explore if machine learning can provide superior solutions.

Jet Flavor Tagging

Jet flavor tagging is one of the cornerstones of the ATLAS physics program. It is concerned with the identification of the jet flavor (bottom, charm or light) based on characteristics such as lifetime, leading to displaced secondary vertices and tracks. We explore how deep learning can improve the performance of these algorithms.

Boosted Jet Tagging

The identification of so-called high-pT boosted jets is becoming more and more relevant as we explore higher energies and increase our sensitivity to higher masses of hypothetical new particles we search for at the LHC. We explore how deep learning can improve the performance of these algorithms.

Next generation pixel-based tracking detectors

Particle trackers based on position-sensitive silicon sensors are at the core of modern collider experiments. High particle intensities and large instrumented areas require new sensor developments. Reconstructing particle tracks from the sensor hits is a challenging combinatorial problem.

We are investigating the full tracking pipeline from sensor hardware to reconstructed physics object. We develop and test novel monolithic silicon sensor based on commercial high-voltage CMOS technologies and we are establishing high-performance reconstruction software to solve the reconstruction challenge.