Ningyuan (Teresa) Huang

Flatiron Research Fellow, Center for Computational Mathematics, Flatiron Institute.

prof_pic.jpeg

I am a Flatiron Research Fellow in the Center for Computational Mathematics at Flatiron Institute. I am broadly interested in geometric deep learning, theory and algorithms for foundation models, and machine learning for science.

I completed my PhD in 2024 from Johns Hopkins University, where I was fortunate to be co-advised by Professor Soledad Villar and Professor Carey Priebe. I was a research intern at Apple Machine Learning Research in 2023 and 2024.

Selected Publications

Graph Machine Learning

  1. Approximately Equivariant Graph Networks
    Ningyuan Huang, Ron Levie, and Soledad Villar
    Advances in Neural Information Processing Systems (NeurIPS), 2023
    PDF Poster
  2. Fine-grained Expressivity of Graph Neural Networks
    Jan Böker, Ron Levie, Ningyuan Huang, Soledad Villar, and Christopher Morris
    Advances in Neural Information Processing Systems (NeurIPS), 2023
    PDF
  3. From Local to Global: Spectral-Inspired Graph Neural Networks
    Ningyuan Huang, Soledad Villar, Carey E. Priebe, Da Zheng, Chengyue Huang, Lin Yang, and Vladimir Braverman
    NeurIPS 2022 Workshop: New Frontiers in Graph Learning, 2022
    PDF

Theory and Algorithms for Foundation Models

  1. Understanding Input Selectivity in Mamba: Impact on Approximation Power, Memorization, and Associative Recall Capacity
    Ningyuan Huang, Miguel Sarabia, Abhinav Moudgil, Pau Rodriguez, Luca Zappella, and Federico Danieli
    International Conference on Machine Learning (ICML), 2025
    PDF Poster
  2. Visualsem: A High-Quality Knowledge Graph for Vision and Language
    Houda Alberts, Ningyuan Huang, Yash Deshpande, Yibo Liu, Kyunghyun Cho, Clara Vania, and Iacer Calixto
    EMNLP First Workshop on Multilingual Representation Learning, 2021
    PDF Code

Machine Learning for Science

  1. CosmoBench: A Multiscale, Multiview, Multitask Cosmology Benchmark for Geometric Deep Learning
    Ningyuan Huang, Richard Stiskalek, Jun-Young Lee, Adrian E Bayer, Charles Margossian, Christian Kragh Jespersen, Lucia A Perez, Lawrence K Saul, and Francisco Villaescusa-Navarro
    Advances in Neural Information Processing Systems (NeurIPS), Datasets and Benchmarks Track, 2025
    PDF Code Poster Website
  2. Diffusion for Fusion: Designing Stellarators with Generative AI
    Misha Padidar, Ningyuan Huang, Andrew Giuliani, and Marina Spivak
    NeurIPS 2025 AI for Science Workshop, 2025
    PDF