PhD Candidate in Machine Learning and Particle Physics

University of California, San Diego

Welcome

I’m Raghav, a PhD candidate at the University of California, San Diego, part of the CMS experiment at the Large Hadron Collider at CERN, and a 2021 Fermilab LPC Artificial Intelligence Fellow.

My focus is on developing machine learning models for particle physics. I’m particularly interested in generative models for simulating particle collisions, as well as more generally using deep learning for improving data collection, reconstruction, and analysis - and ultimately for finding new physics!

I have technical experience in software and electrical engineering, and in the past I have also worked in experimental quantum information science/AMO as well as in the intersection between neuroscience and physics.

Last updated: 28/01/21

Can you find what links the backgrounds?

Interests
  • Generative Models
  • Geometric Deep Learning
  • Di-Higgs Searches
Education
  • PhD in Physics, 2019 -

    University of California, San Diego, 3.97/4.00

  • BSc in Physics and Computer Engineering, 2019

    University of California, San Diego
    summa cum laude, 3.98/4.00

Projects

*
MPGAN

MPGAN

Leading the effort on developing a graph-based generative adversarial network, which we call MPGAN, which has proven effective at generating sparse data with irregular underlying geometry. Our latest work on this has been accepted to the 2021 NeurIPS main conference. We’re now experimenting with a conditional GAN version and variable-sized graphs, as well as applications to other datasets such as CERN detector data.

JetNet

JetNet

Developing a library and website for convenient access to jet datasets, in particle cloud representations, along with several useful utilities for jet-based machine learning development. Has been used so far for our group’s MPGAN and LGAE projects, and continues to be expanded.

Lorentz Group Equivariant Autoencoder

Wrote a review of deep learning models that are equivariant to physics-relevant group transformations for Prof. John McGreevy’s fantastic group theory course. This led to our group developing a graph-based autoencoder equivariant to Lorentz group transformations (LGAE). Preprint on our latest work is in progress.

Particle Graph Autoencoders

Particle Graph Autoencoders

Developing graph-based autoencoders for compression of and anomaly detection in particle cloud representations of Large Hadron Collider data. Our latest work has been accepted to the Machine Learning for Physical Sciences workshop at NeurIPS 2021.

Optical Tweezers and a Quantum Gas Microscope

Optical Tweezers and a Quantum Gas Microscope

Created dynamic, sub-micron holographic optical tweezers and a Quantum Gas Microscope with sub-micron resolution in order to manipulate individual atoms (or qubits) for quantum computing and quantum information science experiments. This work won a William A. Lee Research award, and will be published soon.

Sequential Modeling for Soccer Predictions

Fun project mostly to gain experience with RNNs and Attention. I achieved a 71% testing accuracy in predicting the outcome of European soccer matches.

GRAD: An interactive graph-based degree planning app

GRAD: An interactive graph-based degree planning app

Created an app for visualizing course requirements with a user-friendly UI. I was the Back-end and Algorithms Lead for a team of 10, and personally wrote the server, scraping and graphing algorithms for the app. We were one of 8 finalists out of 60 projects in the UCSD 2018 software engineering course.

Awards

2020-21 Carol and George Lattimer Award for Graduate Excellence
One of two recipients of this award, which ‘honors outstanding graduate students in the Division of Physical Sciences who seek interdisciplinary approaches to problem solving and have a strong commitment to education, mentorship, and service.’
Fermilab LPC Artificial Intelligence Fellowship
For ML-based fast simulation software, ML techniques for reconstruction, compression, and anomaly detection tasks, and a boosted Higgs to WW tagger for precision measurements.
CERN Openlab Summer Students Lightning Talks Award Runner-Up
For the talk ‘Deep Graph Neural Networks for Fast HGCAL Simulation’
2019 IRIS-HEP Fellowship
For the project ‘HGCAL Fast Simulation with Graph Networks’
2019 John Holmes Malmberg Prize
Sole recipient of this prize, which is ‘presented annually at commencement to a graduating physics student who is recognized for potential for a career in physics and a measure of experimental inquisitiveness.’
2018-2019 Physical Sciences Dean’s Undergraduate Award for Excellence
One of 33 students from the departments of Mathematics, Physics and Chemistry ‘recognized for excellence in academics and fundamental research’.
2018 William A. Lee Undergraduate Research Award
For the project ‘Arbitrary ultra-cold atomic lattices using holographic optical tweezers’

Experience

 
 
 
 
 
Machine Learning + Particle Physics Researcher
Sep 2019 – Present San Diego, USA
  • Developing new graph generative models for sparse and irregular data like that prevalent in particle physics
  • Graph neural network (GNN) autoencoders for compression and anomaly detection, machine learning for particle flow reconstruction, Lorentz-group equivariant autoencoders, JetNet library for convenience and reproducibility in machine learning development in high energy physics
  • Developing and applying state-of-the-art GNN classifiers to set the most stringent constraints to date on double-Higgs production, allowing insight into the metastability of the universe
 
 
 
 
 
CERN Openlab Intern
Jun 2019 – Aug 2019 Geneva, Switzerland
  • Started our project on graph generative models for particle physics simulations, motivated primarily by the CMS experiment’s new High Granularity Calorimeter (HGCAL)
 
 
 
 
 
Neurophysics Researcher
Sep 2018 – Jun 2019 San Diego, USA
  • Used two-photon microscopy to measure pO2 in the mouse somatosensory cortex
  • Imaged the cortex to measure vasomotion relative to pO2
 
 
 
 
 
Experimental Quantum Information Science Researcher
Jun 2017 – Jun 2019 San Diego, USA
  • Designed and implemented a setup for a quantum gas microscope (QGM) to image with single-site resolution
  • Generated 2D dynamic, arbitrarily arranged, sub-micron optical tweezers, integrated with the QGM, via two methods, using: 1) a Digital Micromirror Device (i.e. holography), and 2) an acousto-optic deflector
  • Characterized a high (0.8) Numerical Aperture objective for the QGM using OSLO optical simulations and point-spread function image analysis in Python
  • Using an FPGA device, outputted RF waveforms that modulate laser beams with parabolic spatial intensity in order to produce a Bose-Einstein Condensate
  • Programmed FPGA and C electronic devices, and created and (3D) printed mechanical mounts and electronics circuits for experimental use
 
 
 
 
 
Software Intern
Jul 2016 – Sep 2016 Mumbai, India
  • Interned at a software startup which has since been bought by Moka
  • Developed and deployed a location prediction SparkJava server with Cassandra and Redis databases
  • Implemented ML k-means clustering and SVM linear classification algorithms on location data
  • Wrote NodeJS servers and pages for receiving users’ predicted locations and displaying the live data on maps
  • Designed Cassandra and MySQL databases storing user tracking data, and wrote server APIs for accessing/updating, along with web panels for easy viewing of the data (using said APIs)

Contact