I am a physics PhD Candidate at the Massachusetts Institute of Technology (MIT). My background is in particle physics, mathematics, data science, and applied sciences in engineering.
Particle Physics
Quantum Field Theory
Jets and QCD
Future Colliders
Beyond the Standard Model
Data Science
Machine Learning
Optimal Transport
PhD Candidate in Physics, 2020-
Massachusetts Institute of Technology
Advisor: Jesse Thaler
Bachelor of Science in Physics, Mathematics, and Applied Science 2016-2020
Rutgers University
summa cum laude, Highest Honors
Advisor: Stephen Schnetzer
I am primarily interested in the intersection between theoretical particle physics and modern machine learning methods. How can we use machine learning to understand particle physics, and how can we use particle physics to understand machine learning?
We propose "Moment Pooling", a natural extension of Deep Sets networks which drastically decrease latent space dimensionality of these networks while maintaining or even improving performance. Moment Pooling generalizes the summation in Deep Sets to arbitrary multivariate moments, which enables the model to achieve a much higher effective latent dimensionality for a fixed latent dimension. We demonstrate Moment Pooling on the collider physics task of quark/gluon jet classification by extending Energy Flow Networks (EFNs) to Moment EFNs. The smaller latent dimension allows for the internal representation to be directly visualized and interpreted, which in turn enables the learned internal jet representation to be extracted in closed form.
We propose taking advantage of the correlations induced by momentum conservation between jets in order to improve their jet energy calibration. This approach is demonstrated with simulated jets from the CMS Detector, yielding a 3-5% relative improvement in the jet energy resolution, corresponding to a quadrature improvement of approximately 35%.
We demonstrate the utility of muon beam-dump experiments for new physics searches at energies from 10 GeV to 5 TeV. We find that, even at low energies like those accessible at staging or demonstrator facilities, it is possible to probe new regions of parameter space for a variety of generic BSM models.
We show that the Earth Mover's Distance is the natural structure for comparing collider events and constructing IRC-safe observables. We also present SHAPER, a framework for defining and computing shape-based observables, and show how these new observables may be useful for phenomenological studies.
We demonstrate how some recent proposals for both simulation-based and data-based calibrations inherit properties of the sample used for training, which can result in biases for downstream analyses, and argue that our recently proposed Gaussian Ansatz approach can avoid some of the pitfalls of prior dependence.
We present a machine learning framework for performing maximum likelihood inference with Guassian uncertainty estimation, which also quantifies the mutual informaiton between the unobservable and measured quantities.
We search for evidence of Vector-Like Quark (VLQ) production in the CMS Detector of the Large Hadron Collider's entire second run, and set (as of publication) the highest mass limits on VLQ hadronic decays
Below is a list of all public presentations I have given, organized by topic (in no particular order). Note that while several talks share the same category and even the same title, the talks themselves can be very different, especially if the format (invited seminar versus talk, for example) is different! Unfortunately, many slides originally had GIFs and animations, which will not render here as PDFs.
Items marked with a red star (★) are my recommended versions for viewing!
★"Moments Of Clarity in Machine Learning for Jet Physics", Invited Talk, 28 May 2024, for the SLAC AI Seminar. SLAC (Virtual). Slides [PDF]
★ "How Do I Take my Cup of CMS Open Data?", Invited Talk, 11 July 2023, at Fermilab CMS Open Data Workshop 2023. Fermilab, Illinois, United States. Slides [PDF]
★"Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics", Invited Talk, 22 Novemebr 2022, for ATLAS Jet/Etmiss. CERN (Virtual). Slides [PDF]
"Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics", Invited Seminar, 13 September 2022, for the UCI Physics Astro/Particle-ML Seminar series. University of Califoria, Irvine (Virtual). Slides [PDF]
"The New Physics Case for Beam-Dump Experiments with Accelerated Muon Beams", Talk, 16 May 2024, at DPF-Pheno 2024. University of Pittsburgh: Pittsburgh, Pennslyvania. Slides [PDF]
★"The New Physics Case for Muon Beam-Dump Experiments", Internal Seminar, 19 March 2024, for the MIT CTP Graduate Seminar. MIT: Massachusetts, United States. Slides [PDF]
★"SPECTER: Efficient Evaluation of the Spectral EMD ", Talk, 15 August 2024, at the IAIFI Workshop. MIT Media Lab. Slides [PDF]
"SPECTER: Efficient Evaluation of the Spectral EMD ", Talk, 31 July 2024, at BOOST 2024. Palazzo Ducale: Genova, Italy. Slides [PDF]
"SPECTER: Efficient Evaluation of the Spectral EMD ", Talk, 8 November 2023, at ML4Jets2023. DESY: Hamburg, Germany. Slides [PDF]
Can You Hear the Shape of a Jet? An IAIFI Story", Internal Seminar, 16 May 2023, for the IAIFI Engineering Advisory Board. MIT: Massachusetts, United States. Slides [PDF]
★Can You Hear the Shape of a Jet?", Internal Seminar, 17 March 2023, for the MIT CTP Graduate Seminar. MIT: Massachusetts, United States. Slides [PDF]
"Can You Hear the Shape of a Jet?", Talk, 4 Novemebr 2022, at ML4Jets2022. Rutgers University: New Jersey, United States. Slides [PDF]
"Can You Hear the Shape of a Jet?", Talk, 15 August 2022, at BOOST 2022. University of Hamburg: Hamburg, Germany. Slides [PDF]
"Can You Hear the Shape of a Jet? An IAIFI Story", Local Presentation, 11 August 2022, for IAIFI/NSF Review. MIT: Massachusetts, United States. Slides [PDF]
"Can You Hear the Shape of a Jet?", Talk, 10 April 2022, at APS April Meeting 2022. New York City: New York, United States. Slides [PDF]
★"Moment Pooling: Gaining Performance and Interpretability Through Physics Inspired Product Structures", Talk, 2 August 2023, at BOOST 2023. LBNL, Berkeley, California, United States. Slides [PDF]
"Moment Pooling: Gaining Performance and Interpretability Through Physics Inspired Product Structures", Talk, 10 April 2023, at APS April Meeting 2023. Minneapolis, Minesota, United States. Slides [PDF]
"Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics", Poster, 3 December 2022, at Machine Learning and the Physical Sciences. NeurIPS (Virtual). Poster [PNG]
"Learning Uncertainties the Frequentist Way: Calibration and Correlation in High Energy Physics", Talk, 7 July 2021, at ML4Jets2021. University of Heidelberg (Virtual). Slides [PDF]