My Passion

driving innovative research at the intersection of physics and machine learning

My primary research interest is to integrate machine learning models with physics knowledge. More generally, I am interested in developing novel solutions to problems involving physical processes through the use of machine learning and probabilistic models.

I am a Ph.D. student at NYU researching experimental particle physics with a machine learning focus under the advising of Kyle Cranmer. In particular, I am working on realizing original applications for machine learning in particle physics analyses. Examples of such research by others include QCD-aware neural networks, using convolutional neural networks for jet reconstruction through jet images, and using generative adversarial networks for efficient simulations. As an undergraduate at Notre Dame, I double-majored in Physics and Honors Mathematics with a concentration in Advanced Physics and graduated with honors in each.

I want to know how God created this world. I'm not interested in this or that phenomenon, in the spectrum of this or that element. I want to know His thoughts; the rest are details.

- Albert Einstein

Physics Interests

Particle Physics
Quantum Computing
Astrophysics

Machine Learning Interests

Simulation-based Inference
Normalizing Flows
Generative Models
Bayesian Methods
Reinforcement Learning

Personal Interests

Gastronome Explorer Hiker Fitness Enthusiast Golden Domer Space Advocate Volunteer

Lorentz Flow for Jets

a normalizing flow model for boosted jets

We developed a model for the distribution of three-momenta of particles constituting the two boosted jets in dijet production. We start with a simple rest frame model and then use the Lorentz transformation to transform this into a model of boosted momenta conditioned on the velocity. Maximum likelihood estimation can then quickly estimate this velocity for a dataset. In addition, this model has potential for efficiently generating jet data, tagging jets, or inferring other physical parameters.


Ginkgo + Cluster Trellis

tuning parton shower parameters with the marginal likelihood

Tuning parton shower models to data is an important task for HEP experiments. We are performing exploratory research for what tuning the parton shower might look like if the parton shower were described by a generative model with a tractable likelihood. For this work we consider the Ginkgo model, which is a simplified parton shower that has been designed to facilitate this research. Ideally, we would tune the model with a maximum likelihood fit. The challenge is that the likelihood for the data given the model parameters requires marginalizing over the (2N-3)!! possible showering histories, where N is the number of jet constituents. We demonstrate that with the hierarchical cluster trellis we can exactly marginalize over this enormous space of showering histories and fit the parameters of the Ginkgo model.

there is a flower
whose colour I cannot see
of pervasive scent
the name
of the end of all things, in all things .

-John Riley