Elyssa Hofgard

  • Program Year: 1
  • Academic Institution: Massachusetts Institute of Technology
  • Field of Study: Electrical Engineering and Computer Science
  • Academic Advisor: Tess Smidt
  • Practicum(s): Practicum Not Yet Completed
  • Degree(s):
    M.S. Computational and Mathematical Engineering, Stanford University, 2022; B.S. Physics, Stanford University, 2021

Summary of Research

I'm developing symmetry equivariant neural networks (ENNs) for applications in the natural sciences. If our data originates from or describes a physical system, we know it's subject to physical laws which have known symmetries. We can achieve superior data-efficiency, generalization, and robustness to physical laws when using ENNs as opposed to standard neural networks through incorporating physical symmetries. Currently, I'm working on using the mathematics of symmetry-breaking and properties of ENNs to uncover symmetry-implied missing information from diverse physical data. These methods will be broadly applicable to anomaly detection-identifying when our observations of physical systems imply we do not have a complete description and determining how the unobserved information must transform by symmetry (e.g. resolve galactic dynamics caused by unobservable black holes or dark matter). I aim to build generalized algorithms that aid in the understanding and engineering of physical systems.


Wang, R., Hofgard, E., H. Gao, R. Walters, and T. Smidt. "Discovering Symmetry Breaking in Physical Systems with Relaxed Group Convolution," Feb 2024, https://arxiv.org/abs/2310.02299.

"Artifical Intelligence for Science in Quantum, Atomistic, and Continuum Systems," Nov 2023, https://arxiv.org/abs/2307.08423.

Cotter, D., Hofgard, E., Szpiech, A., & Rosenberg, N. "A rarefaction approach for measuring population differences in rare and common variation," April 2023. Accepted, Genetics.

ATLAS Collaboration. "Search for Dark Photons from Higgs Boson Decays via ZH Production with a Photon plus Missing Transverse Momentum Signature from pp Collisions at s = 13 TeV with the ATLAS Detector," December 19, 2022. https://doi.org/10.48550/arXiv.2212.09649.

Garg, Rocky Bala, Elyssa Hofgard, Lauren Tompkins, and Heather Gray. "Exploration of Different Parameter Optimization Algorithms within the Context of ACTS Software Framework," November 1, 2022. https://doi.org/10.48550/arXiv.2211.00764.

Hofgard, Elyssa. "Cuts Optimization and Machine Learning Models for Dark
Photon Signal-Background Discrimination with the ATLAS Detector," June 2021. Stanford Digital Repository, Undergraduate Theses. purl.stanford.edu/pb798fj0435.

Hofgard, Elyssa F., Garry Chinn, and Craig S. Levin. "Simultaneous Multi-Isotope PET: A Computational Framework for Line of Response (LOR) Identification." In 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference (NSS/MIC), 2020. https://doi.org/10.1109/NSS/MIC42677.2020.9507891. Oral presentation also given at 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference.

"GWTC-2: Compact Binary Coalescences Observed by LIGO and Virgo during the First Half of the Third Observing Run." Physical Review X 11, no. 2 (June 9, 2021): 021053. https://doi.org/10.1103/PhysRevX.11.021053. As a member of the LIGO Scientific Collaboration (LSC), I met the contribution criteria for inclusion on the LSC author list, and I earned authorship on eleven LSC publications. See my Google Scholar page for the full list (https://scholar.google.com/


QuARC (MIT Center for Quantum Science and Engineering Annual Conference), Jan 2024, Best Poster in Quantum Algorithms and Machine Learning Session.

Recipient, Department of Energy Computational Science Graduate Fellowship (DOE CSGF), April 2023.

Recipient, Department of Defense National Defense Science and Engineering Graduate Fellowship (NDSEG), April 2023. Declined for DOE CSGF.

Recipient, MIT Advanced Television and Signal Processing Fellowship, 2022-2023 academic year.

NSF GRFP Honorable Mention, April 2022.

Firestone Medal for Excellence in Undergraduate Research, August 2021. Award given to the top ten percent of all honors theses in the social sciences, natural sciences, and engineering and applied sciences.

American Physical Society Far West Section Meeting, Nov 2019, Honorable Mention for Best Undergraduate Poster.