Duy-Tung Pham

Hello! I’m Duy-Tung Pham (/zwi-tuŋ fam/). In Vietnamese, my name is written as Phạm Duy Tùng. You can call me Duy-Tung or just Tung for short!

Currently, I’m an AI Research Resident at FPT Software AI Center, where I have the great opportunity to work under the guidance of Prof. Thieu Vo and Prof. Tan Minh Nguyen.

Before this, I completed my Bachelor’s degree in Computer Science in Honors Program at
School of Information and Communication Technology (SoICT), Hanoi University of Science and Technology (HUST). During my studies, I was also a research student at the Data Science Laboratory, mentored by Dr. Linh Ngo Van and Prof. Khoat Than.

(A little tip: you can switch between dark and light mode by clicking the button on the right of the CV tab!)

Research Interest

My current research interests lie in the mathematical foundations of deep learning, with an emphasis on geometric and dynamical perspectives for designing efficient, scalable, and robust models from first principles.

(Looking for PhD in CS/ML) I am seeking PhD positions in Computer Science and Machine Learning to advance research and innovation in the field.

Publications

(*) denotes equal contribution, (**) denotes equal advising.

Dynamical Properties of Tokens in Self-Attention and Effects of Positional Encoding

Published in The Thirty-ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025)
Duy-Tung Pham*, An Nguyen The*, Viet-Hoang Tran, Nhan-Phu Chung, Xin T. Tong, Tan Minh Nguyen**, Thieu Vo**

NeuroMax: Enhancing Neural Topic Modeling via Maximizing Mutual Information and Group Topic Regularization

Published in Findings of the Association for Computational Linguistics: EMNLP 2024
Duy-Tung Pham*, Trang Nguyen*, Tung Nguyen*, Linh Ngo Van, Duc Anh Nguyen, Thien Huu Nguyen

Demystifying the Token Dynamics of Deep Selective State Space Models

Published in The Thirteenth International Conference on Learning Representations (ICLR 2025), Spotlight, top 5.1%
Thieu Vo, Duy-Tung Pham, Xin T. Tong**, Tan Minh Nguyen**

Out-of-vocabulary handling and topic quality control strategies in streaming topic models

Published in Neurocomputing, 2025
Tung Nguyen*, Duy-Tung Pham*, Linh Ngo Van, Ha-Bang Ban, Khoat Than

Equivariant Neural Functional Networks for Transformers

Published in The Thirteenth International Conference on Learning Representations (ICLR 2025)
Viet-Hoang Tran*, Thieu Vo*, An Nguyen The*, Tho Tran Huu, Minh-Khoi Nguyen-Nhat, Thanh Tran, Duy-Tung Pham, Tan Minh Nguyen

TopiCOT: Neural topic model aligning with pre-trained clustering and optimal transport

Published in Neurocomputing, 2025
Tung Nguyen*, Duy-Tung Pham*, Quang Duc Nguyen, Linh Ngo Van, Anh Nguyen Duc, Sang Dinh Viet

Equivariant Polynomial Functional Networks

Published in Forty-second International Conference on Machine Learning (ICML 2025)
Thieu Vo*, Viet-Hoang Tran*, Tho Tran Huu, An Nguyen The, Thanh Tran, Minh-Khoi Nguyen-Nhat, Duy-Tung Pham, Tan Minh Nguyen

Preprints

Mixed-Curvature Tree-Sliced Wasserstein Distance

Under review
Duy-Tung Pham*, Viet-Hoang Tran*, Thieu Vo, Tan Minh Nguyen