Dung V. Nguyen

About Me

I am a master student who is passionate about AI and its transformative potential. My interests lie in mathematics and its applications, particularly in optimization and deep learning. I aspire to an academic career, contributing to both the theoretical and practical aspects of mathematics in AI.

My research focuses on uncovering the mathematical intuition behind AI models and leveraging mathematical tools to optimize models and enhance computational efficiency. I seek to collaborate with faculty and researchers who share my enthusiasm for advancing mathematical knowledge and developing innovative solutions to complex problems.

My Thought on Research Approach

Throughout my academic research experience, I have identified recurring patterns in methodology that appear in many papers. This observation has led me to categorize AI research skills into three distinct levels:

Level 1 ("LEGO Assembly"): Researchers naively combine modules from different architectures, relying on trial and error to improve performance.

Level 2 (Methodological Abstraction): Researchers integrate techniques to offset weaknesses, uncovering hidden connections and introducing novelty.

Level 3 (Mathematical Modeling): Researchers transform problems into mathematical frameworks, ensuring scalability, interpretability, and robustness for future advancements.

Research Interests

Mixture of Experts Optimization Computer Vision Reinforcement Learning

Education

Master of Science in Mathematics and Informatics

Hanoi University of Science and Technology

2024 - Present

Relevant coursework: Advanced Machine Learning, Optimization Methods, Computer Vision, Reinforcement Learning

Bachelor of Science in Mathematics and Informatics

Hanoi University of Science and Technology

2019 - 2023

Graduated with Honors

Selected Publications

• CAMEx: Curvature-aware Merging of Experts

  International Conference on Learning Representations (ICLR), 2025

  Viet Dung Nguyen, Minh Nguyen Hoang, Luc Nguyen, Rachel Teo, Tan Minh Nguyen, Linh Duy Tran

  View Publication PDF

• A Hyper-Transformer model for Controllable Pareto Front Learning with Split Feasibility Constraints

  Neural Networks, 2024

  Tran Anh Tuan, Nguyen Viet Dung, Tran Ngoc Thang

  View Publication PDF

• A neurodynamic approach for a class of pseudoconvex semivectorial bilevel optimization problems

  Optimization Methods and Software, 2024

  Tran Ngoc Thang, Dao Minh Hoang, Nguyen Viet Dung

  View Publication PDF

• An EffcientNet-encoder U-Net Joint Residual Refinement Module with Tversky–Kahneman Baroni–Urbani–Buser loss for biomedical image Segmentation

  Biomedical Signal Processing and Control, 2023

  Do-Hai-Ninh Nham, Minh-Nhat Trinh, Viet-Dung Nguyen, Van-Truong Pham, Thi-Thao Tran

  View Publication PDF

Contact

Blogs

• Mixture of Experts: A Comprehensive Overview

  In this blog, I explore the concept of Mixture of Experts (MoE), its mathematical foundations, and its applications in deep learning. I also discuss how MoE can be used to improve model performance and scalability.

  Read More

• Multi-Agent Reinforcement Learning: A Comprehensive Overview

  In this blog, I explore the concept of Multi-Agent Reinforcement Learning (MARL) (on going)

  Read More