By Deborah Acheampong

Can machines assist human creativity? SFU computing science professor Richard Zhang has been excited about creative modeling and, with his team of collaborators, he has been answering the fundamental question of whether machines can learn to be creative or, to a lesser extent, assist in pursuing creativity when humans remain in the loop. In earlier works, they applied concepts from evolutionary biology to develop the first 3D modeling tool that can produce creative outcomes. Then, they explored collaborative 3D modeling in the context of co-creativity. Most recently, they made strides in creative 2D content generation, specifically artistic typography, by tapping into the power of diffusion and large language models.

As a Visual Computing expert, Zhang specializes in computer graphics, 3D computer vision, and geometric deep learning. His work encompasses all scientific and computational disciplines that acquire, analyze, and generate visual data such as images, 3D shapes, and virtual environments.

Zhang's research has received several prestigious awards, including the 2020 Best Student Paper Award at the Computer Vision and Pattern Recognition (CVPR), one of the most prestigious conferences in visual computing and machine learning, and the Computer Graphics Achievement Award from the Canadian Human-Computer Communications Society (CHCCS) in 2021. In 2020, Zhang was recognized as a Distinguished SFU Professor for his outstanding performance and achievements and his international pre-eminence in visual computing. While his expertise is hugely relevant to the industry, Zhang’s academic contributions are also significant, publishing more than 150 papers, mainly in the top venues of his field, and with a total citation over 18,000 times.

Professor Zhang obtained his BSc and MSc in mathematics from the University of Waterloo and his PhD from the University of Toronto, all in Canada.

Recognizing his outstanding research record driving academic and industry standards, Zhang has been named a 2024 Institute of Electrical and Electronics Engineers (IEEE) Fellow for his “contributions to shape analysis and synthesis in visual computing.” IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. It has more than 450,000 members in more than 190 countries and is a leading authority in many areas, including engineering, computing, and technology information. Less than one-tenth of one percent of IEEE members worldwide are selected as Fellows in any year. Fellow status is awarded to individuals with "an outstanding record of accomplishments in any of the IEEE fields of interest."

We spoke to Richard Zhang about his research and his stellar contributions to the visual computing field.

Tell us about your contributions to visual computing that led to your recognition as an IEEE Fellow.

My contributions to visual computing are in shape analysis and synthesis. Shape analysis helps a machine to understand the 3D world. If a robot is to clean our houses, it needs to recognize the household items and how to interact with them to perform the task.

We see many social media coverages on generative AI these days. Shape synthesis is at the heart of this development, focusing on 3D content creations for games, virtual and augmented reality (VR/AR), computer-aided design and manufacturing. My research group at SFU and my collaborators worldwide have developed some of the most advanced shape analysis and synthesis technologies to serve these and many other applications.

What emerging trends in visual computing do you find particularly exciting or promising for the future?

It is the emergence of large foundational models, including language and vision-language models, and powerful generators, such as diffusion models. They have generated a great deal of excitement in generative AI and produced results of unprecedented quality, especially in images and video. 3D generative AI is also flourishing but still faces many challenges.

Where do you see promising areas for future research in shape analysis and synthesis?

I foresee 3D generative AI as a real challenge; we are not yet there. There is still very little 3D data to train machine learning algorithms. The current trend is to avoid the use of 3D supervision in training machines to perform 3D reasoning, but this is not only unnatural (babies learn the 3D world by interacting with it) but also comes with a huge cost in terms of training time, network size, and even environmental impact. A promising direction is to think of how to judiciously select good 3D data, even synthesizing them, so that small 3D data combined with smart representations and training schemes can outperform the very large models. I am also a strong proponent of the “learning by interacting” paradigm, which combines computer vision, graphics, and robotics to cross-fertilize these fields. This is an area where SFU visual computing has a great deal of research strength.

What does your recognition as a 2024 IEEE Fellow mean to you personally and professionally?

Such recognition by my peers is an honor and brings some personal satisfaction. I also joked with some friends that it feels like one of those perks you get when you age. It recognizes efforts from all my students, postdocs, and collaborators. I owe a great deal to them and SFU for fostering my academic career. As a fellow myself now, it means that I will be in a better position to nominate my deserving colleagues to recognize their accomplishments.