Hien Dang

Hawaii_profile_pic_2.png

Welcome! My name is Hien and I am an incoming PhD student at Department of Statistics and Data Science, University of Texas at Austin. Before that, I was a research resident at FPT Software AI Center. I graduated with honor bachelor degree from the Department of Mathematics and Computer Science, Ho Chi Minh University of Science, Vietnam National University.

Email: danghoanghien1123@gmail.com

Research Interests

My research interests are in the theory and mathematical foundation of Deep Learning. Specifically, my recent works focus on understanding “collapse” phenomena observed in deep learning architectures such as neural collapse in deep overparameterized neural networks or posterior collapse in variational autoencoders.

“*” and “**” denote equal contribution

Recent news

Jan 1, 2024 Our paper “Beyond Vanilla Variational Autoencoders: Detecting Posterior Collapse in Conditional and Hierarchical Variational Autoencoders” has been accepted as a poster at ICLR 2024.

Publications

  1. Arxiv
    Neural Collapse for Cross-entropy Class-Imbalanced Learning with Unconstrained ReLU Feature Model
    Hien Dang, Tho Tran, Tan Nguyen**, and Nhat Ho**
    Under review, 2024
    HTML PDF
  2. ICLR
    Beyond Vanilla Variational Autoencoders: Detecting Posterior Collapse in Conditional and Hierarchical Variational Autoencoders
    Hien Dang, Tho Tran, Tan Nguyen**, and Nhat Ho**
    12th International Conference on Learning Representation, 2024
    HTML PDF
  3. ICML
    Neural Collapse in Deep Linear Networks: From Balanced to Imbalanced Data
    Hien Dang*, Tho Tran*, Stanley Osher, Hung Tran-The, Nhat Ho**, and Tan Nguyen**
    Proceedings of the 40th International Conference on Machine Learning, 2023
    HTML PDF