Research Content

During my PhD, I worked on the development of a new search strategy for New Physics in ATLAS data. The objective of this strategy is to perform a model independent search that requires no signal hypothesis and as few data simulation as possible.

For this purpose, I developed an anomaly detection algorithm based on unsupervised Machine Learning techniques. The objective was to be able to distinguish a potential signal from the Standard Model background without making any assumption on the New Physics models.

In order to identify any deviation of the data in a model independent way, I implemented a improved version of the BumpHunter algorithm, as well as a data-driven background estimation procedure.In parallel of this work, I also work on the maintenance and monitoring of the ATLAS Tile calorimeter.

For my postdoctoral fellowship, I work on developing online anomaly detection models based on fast Machine Learning to be used as advanced trigger for ATLAS detector. I also work on the new iTk pixel detector that will be deployed for the High-Luminosity upgrade of ATLAS.

Related Research Results

Website