Elyssa Hofgard

Cambridge, MA

I’m currently a PhD student in Electrical Engineering and Computer Science (EECS) at MIT. I’m a proud member of the Atomic Architects Research Group, advised by Professor Tess Smidt. I also work closely with the Griffin Group in the Molecular Foundry at Lawrence Berkeley National Laboratory. I received a master’s degree in Computational and Mathematical Engineering (‘22) and a bachelor’s degree in physics with honors (‘21), both from Stanford University. I’m funded by the Department of Energy Computational Science Graduate Fellowship.

I worked with the ATLAS collaboration at CERN during undergrad and my master’s degree with Professor Lauren Tompkins. In my PhD, I pivoted to computational materials science and machine learning (ML) development. I’m passionate about applying computational tools to materials design and problems in condensed matter physics. I enjoy working on projects where I am able to collaborate across disciplines, learn about a new scientific field, and interface with experimentalists. Lately, I have been working on modeling phase transitions in exotic magnetic materials, characterizing length scales of order in amorphous materials, and developing representation learning methods for applications in powder X-ray diffraction. See my Google Scholar and my CV for more information.

Outside of research, I love to hike, backpack, and spend time outdoors. I also am an avid swimmer and a member of the MIT Women’s Chorale.

If you’re interested in probability theory or PDEs, you may be looking for my brother Jake Hofgard. He is a PhD student in mathematics at UC Berkeley and an excellent mountaineer.

news

Mar 16, 2026	Presenting Using Equivariant Neural Networks to Learn Multipolar Order Parameters at the APS Global Physics Summit 2026 — Altermagnetism II session, Mon. March 16, 3:30–6:18 p.m., Convention Center Room 503. Come say hi if you’re at APS!
Jan 26, 2026	Our paper To Augment or Not to Augment? Diagnosing Distributional Symmetry Breaking has been accepted at ICLR 2026!

latest posts

Mar 02, 2026	To Augment or Not to Augment? Diagnosing Distributional Symmetry Breaking

selected publications

To Augment or Not to Augment? Diagnosing Distributional Symmetry Breaking

Hannah Lawrence^*, Elyssa Hofgard^*, Vasco Portilheiro, and 3 more authors

In International Conference on Learning Representations, 2026

arXiv Bib HTML

@inproceedings{lawrence2025augment,
  title = {To Augment or Not to Augment? Diagnosing Distributional Symmetry Breaking},
  author = {Lawrence, Hannah and Hofgard, Elyssa and Portilheiro, Vasco and Chen, Yuxuan and Smidt, Tess and Walters, Robin},
  year = {2026},
  booktitle = {International Conference on Learning Representations},
}

Relaxed Equivariant Graph Neural Networks

Elyssa Hofgard, Rui Wang, Robin Walters, and 1 more author

In ICML 2024 Workshop on Geometry-grounded Representation Learning and Generative Modeling, 2024

arXiv Bib

@inproceedings{hofgard2024relaxed,
  title = {Relaxed Equivariant Graph Neural Networks},
  author = {Hofgard, Elyssa and Wang, Rui and Walters, Robin and Smidt, Tess},
  booktitle = {ICML 2024 Workshop on Geometry-grounded Representation Learning and Generative Modeling},
  year = {2024},
}

Discovering Symmetry Breaking in Physical Systems with Relaxed Group Convolution

Rui Wang, Elyssa Hofgard, Han Gao, and 2 more authors

In Forty-First International Conference on Machine Learning, 2024

Bib HTML

@inproceedings{wang2024discovering,
  title = {Discovering Symmetry Breaking in Physical Systems with Relaxed Group Convolution},
  author = {Wang, Rui and Hofgard, Elyssa and Gao, Han and Walters, Robin and Smidt, Tess},
  booktitle = {Forty-First International Conference on Machine Learning},
  year = {2024},
}