Personal Website of Ziye(Gaven) Ma
Hi, my name is Ziye Ma (马梓业), and I’m currently an assistant professor in the CS department at City Universtiy of Hong Kong. I am especially grateful to receive the honorary title of presidential assistant professor at CityU. I’m also known as Gaven, so feel free to call me whichever way you prefer. I obtained my Ph.D. from the EECS department of UC Berkeley, supervised by Somayeh Sojoudi. Prior to that, I studied Engineering Science at the University of Toronto and did research at STARS lab. The majority of my early life was spent jumping back and forth between Beijing and Toronto, both of which places I call home.
MOFA lab
I am currently the Principle Investigator of the Machine Optimization and Fundamental Algorithms lab, a place where vibrant young researchers work together to advance the state-of-the-art of ML theory and practices. We have multiple fully-funded PhD and post-doc positions open. Please refer to this page for more detailed information. Our lab is supported through research grants by Research Grant Council of Hong Kong, Natural Science Foundation of China, CityU HK, and other partners.
I also hold an adjuct positon at Shenzhen Loop Area Institue (SLAI), so you’re welcome to apply to my lab through SLAI as well.
Research Interest
My research aims to develop explainable and efficient machine learning systems, emphasizing the exploration of new theoretical tools and perspectives. These efforts are directed towards demystifying various phenomena in modern machine learning. Mathematically, my work predominantly addresses non-convex optimization problems, drawing inspirations from matrix theory, algebraic geometry and other fields.
From a high level, my research mostly involves two parts:
- Machine learning theory, studied under the framework of non-convex landscape analysis. We aim to understand under what conditions ML practitioners can find global or generalizable solutions, and what guarantees accompany them.
- Efficient machine learning practices, guided by theories developed in the first part. We aim to tackle three aspects of efficient machine learning: data efficiency (using less data to achieve similar performance, given the current lack of data), algorithm efficieny (using zeroth order optimization to reduce memory and compute demands for training and fine-tuning), model efficiency (better quantization strategies for compressing model sizes).
For any inquiries, concerns, or suggestions regarding my research, please do not hesitate to contact me via email. I am committed to engaging in fruitful discussions and am always open to feedback.
Fun Facts
I enjoy photography, watching Anime, and experiencing a city by wandering about (a.k.a city walk lol). Photo credits to my Mom, the shot was taken near 观妙亭 of 景山公园 in May of 2021, and in the background is the breathtaking Forbidden City.