Heinz, Michael | Engineering Papers

DOE OSTI2020

Review of "ML for Surface Complexation Model Development"

The presentation given by Jadallah Zouabe on “ML for Surface Complexation Model Development”was extremely informative, but lacked some explanation of details that made the latter half hard to follow as someone outside of the field. He began his presentation with an anecdote to spilling milk or ink on the floor and seeing it spread, cleverly connecting that to the spread of contamination in the real world. He then explained the importance of being able to model contamination, as it allows specialists to see the potential effects on different areas and environments in the coming years. Heal so made it clear that the main issue trying to be addressed here is how expensive modelling is and that their goal is to make computationally cheaper models. The introduction was overall very well formulated and attention-grabbing, but it seemed to take up too much of the presentation time, as after this point the presentation lacked details and explanations.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2020

Improving High-Energy Particle Detectors with Machine Learning

Microseconds after the Big Bang, the universe existed in a state called the quark-gluon plasma (QGP). To experimentally study its properties, the QGP is recreated in high-energy nuclear collisions at the LHC, and the particles produced from the QGP are reconstructed from their energy deposition in the ATLAS calorimeter. This requires both classifying the particles and calibrating their deposited energy. The objective of this project is to improve the reconstruction by using machine learning techniques, where the energy depositions of clusters of cells, formed by ATLAS topo-clustering methods, are treated as three-dimensional images when inputted to neural networks. This approach significantly improves the calibration of deposited energies when cross-validating while training, and models trained on idealized data predict the calibrated energies of particles in more complex data sets well. Additionally, implementation of a data generator using uproot allows the program to load input data into memory as needed while training or predicting, significantly reducing the amount of memory used. The data generator also allows for use of multiprocessing to speed up training and evaluating. This work illustrates that using machine learning methods for both classification and calibration has the potential to significantly improve particle reconstruction.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗