This project is connected to the PhD thesis described here. During the M2 internship, the student will develop the first part of this new HESS analysis strategy: the AI filter called FiBER (Filter Before Event Reconstruction).

The AI filter, based on Unsupervised Graph Neural Network Clustering, optimized on real data, aims to reduce the background contamination from proton showers just after the calibration procedures, with the goal of minimizing the number of gamma-like events to be considered as input in the following shower reconstruction procedure (goal: speed-up of analysis). Ideally, only 10-20% of the whole dataset should be the input of the subsequent shower reconstruction step, but even a 50% reduction from the whole sample will be considered acceptable. The filtering approach will also use modern data visualization techniques as T-SNE, UMAP, PCA and data clustering algorithms as hdbscan.