Woojoo Na

I am a PhD student at Northeastern University, advised by Professor Jennifer Dy. My research interests lie at the intersection of mathematics and machine learning. I am interested in identifying and utilizing algebraic and geometric properties in deep learning and optimization methods.

Previously, I received my B.A. in Mathematics from the University of Oxford and my M.S. in Computer Science from Tufts University.

Before Northeastern, I was fortunate to work with Professor Thomas Lukasiewicz at Oxford Computer Science, Dr. Andrey Kormilitzin at Oxford Mathematical Institute, Professor Christophe Petit at Oxford Mathematical Institute, and Professor Abiy Tasissa at Tufts University.

Email  /  CV  /  Google Scholar  /  GitHub

• 05/2026: ITSPACE was accepted at ICML 2026! [OpenReview]
• 09/2024: Joined Northeastern University as a PhD student.
• 12/2023: Graduated from Tufts University with an M.S. in Computer Science, following the academic retirement of my former supervisor.
• 07/2023: My paper, "RACH-Space: Reconstructing Adaptive Convex Hull Space with Applications in Weak Supervision", is now available.
• 09/2021: Joined Tufts University as a PhD student.
• 08/2021: Discharged from the Korean Army as Sergeant after completing 18 months of service.
• 02/2020: Joined the Korean Army to serve as FDC in artillery.
• 01/2020: I presented our work on cryptography at CMS-Ewha Womans University.
• 12/2019: We presented at 17th IMA International Conference on Cryptography and Coding.
• 07/2019: Graduated from the University of Oxford with a B.A. (Hons) in Mathematics.

Deep learning & Optimization

ITSPACE develops a structure-preserving approach to low-rank covariance transport, connecting principled machine learning with geometry, optimization, and mathematical structure.

Developed a novel algorithm which labels data when given partial, noisy information about the ground-truth labels. State-of-the-art performance on real-world benchmark data.

We focused on the framework for CNNs that are equivariant under arbitrary group transformations, and looked at its applications and limitations, experimenting on the MNIST dataset.

Natural Language Processing

Devised an online platform for fake news detection using neural models, aimed at small investors prone to fake news. Won Brigade Commander Award for Excellence.

Worked on conceptual space of a transformer encoder, studying the similarities between NLP architectures and the human brain.

In this challenge, we proposed a model for detecting Adverse Drug Events and Medication Extraction in EHRs. We exploited CNN architecture and a character-level word embedding model to demonstrate a robust approach to concept extraction and relation extraction.

Worked on knowledge graph completion for ontology reasoning.

Cryptography

By leveraging group properties in Cayley hash functions, we introduce trapdoor attacks on Cayley hash function parameters.

Source code credit to Dr. Barron