Search This Site

Chris Neu - Experimental particle physics, machine learning, data science

Science & Engineering
Supervising Faculty Member: 
Chris Neu
Research Focus: 

My work is done at the Large Hadron Collider, and I study new or exotic forms of matter in order to understand the tiniest building blocks that the world is made of. The qualities and interactions of these building blocks have huge implications on our understanding of the history, condition and fate of the universe. This is basic research, all in the scientific pursuit of understanding everything we can about the world around us.

Basic methodologies involve particle detection apparatus, analysis of electrical signals from these devices, digitization of these signals for processing/storage, and then offline analysis of this data. There are also the generation and study of simulated data samples as well.

Offline analysis often touches on the task of extracting a small or obscured signal among a great deal of mundane, similar-looking backgrounds. This can entail simple signal extraction techniques -- like identifying effective discriminants by hand -- or can involve more intricate approaches, such as machine learning (ML) for signal and background classification.

Position Description: 

The ideal student for this position is someone who has some computing background and is interested in machine learning, specifically applying ML techniques to a large-data problem in experimental physics.

The task will be to build a suite of classifiers that distinguish different categories of events at the Large Hadron Collider -- for instance, identifying Higgs boson decays among a huge swamp of other more mundane particle decay processes.

The student will work with simulated data to understand the differences between signal and background processes. The student will use these simulated data to train these new powerful classifiers. Then the classifiers will be pitted against one another to determine which is optimal under a certain set of conditions.

The student will have the chance to join a vibrant research group with a faculty member, a postdoc, 2-3 PhD students and several other undergraduates.

Required Skills: 

Must be competent in either C++ or python. Familiarity with modern OS's is a must. Fluency in the Linux OS is a plus. Experience with machine learning is a plus. It would be good to be familiar with machine learning concepts in general, and particularly deep neural networks, adversarial neural networks, convolutional neural networks, boosted decision trees, others.

Nothing formal. Good grades in more than one programming class is a benefit.
What will you learn: 

1. To learn about important classification problems in particle physics.
2. To be able to deploy ML techniques for feature classification problems.
3. To be able to appreciate the differences between various techniques and assess their relative performance in a given use case.

Experimental particle physics, machine learning, data science