Unsupervised clustering for collider physics

We propose a new method for unsupervised clustering for collider physics named UCluster, where information in the embedding space created by a neural network is used to categorize collision events into different clusters that share similar properties. We show how this method can be developed into an unsupervised multiclass classification of different processes and applied in the anomaly detection of events to search for new physics phenomena at colliders.

Automated summary

The UCluster method uses information in the embedding space created by a neural network to categorize collision events into different clusters, enabling unsupervised multiclass classification for anomaly detection in collider physics.