research student

Danielle Turvill

Postgraduate Research Student


College of Science and Engineering





My research focuses on applying machine learning to aid particle identification (PID) within ALICE (A Large Ion Colliding Experiment), CERN. In particular, I am using graph learning to allow for more meaningful classifications of physics data.

ALICE was designed to investigate quark-gluon plasma (QGP), a state of matter which occurred shortly after the Big Bang. PID is crucial for QGP measurements.

Generally, PID analyses isolate interesting signals from high levels of combinatorial background. Machine learning is an increasingly popular solution for optimising signal-to-background ratios. However, these methods learn non-trivial relations between physics parameters and so it is not always possible to know exactly what the machine has really learned.

I became interested in creating graph representations of particle interactions as they would allow for the application of various graph-computing techniques and, more importantly, add intelligence over learned data.

The transparency of learning methods is very important within the physics community. By applying learning methods to these graph representations, I aim to provide physicists the opportunity to be able to deduce the validity of interesting physics with more ease, through a more informative and simplified interpretation of relationships between particles.

Thesis title

Investigation of Graph Learning for the analysis of High Energy Physics data from LHC Experiments