Optimization for Deep Learning (OptDL) Lab

Introduction

• OptDL Lab is dedicated to advancing the theoretical foundations and practical methodologies of optimization in modern machine learning. Our research lies at the intersection of optimization, deep learning, and AI, aiming to understand and improve the training, generalization, and interpretability of deep learning and AI models.

Members

• Principal Investigator: Peng Wang

• PhD Students:

  • Yutao Chen: BUPT (B.S.), ANU (M.S.), UM (Ph.D candidate, 2026 Aug - Present)

  • Wendong Li: HUST & U. Birmingham (B.S.), U. Bonn (M.S.), UM (Ph.D candidate, 2026 Aug - Present)

  • Tianrun Zhao: UM (B.S.), NTU (M.S.), UM (Ph.D candidate, 2026 Aug - Present)

• Research Assistants

  • Hanzhang Li: CUFE (B.S.), NUS (M.S.)

Supported Grants

• National Natural Science Foundation of China (NSFC), Young Scientists Fund-Type C, PI, On the Loss Landscape and Error Bounds of Deep Neural Networks, NSFC 12601585 (Jan 2027 - Dec 2029)

• Macau Science and Technology Development Fund (FDCT), Funding Scheme for Innovation and Technology Promotion, PI, Toward Theoretical Foundations of Diffusion Models: Principles from Optimization to Generalization, FDCT/0091/2025/ITP2 (Dec 2025 - Dec 2027)

• Technology and Innovation Start-up Funding Scheme, University of Macau (Dec 2025 - Dec 2027)

• Start-up Research Grant, University of Macau (Aug 2025 - Aug 2027)