On April 22, the 2021 Facebook Fellows were announced, including Cheng Zhang, a computer science Ph.D. student in the Donald Bren School of Information and Computer Sciences (ICS). “It is a great honor for me to receive the Facebook Fellowship this year,” says Zhang. “It is delightful that my research is being recognized and awarded by the industry.”

The Facebook Fellowship program supports promising doctoral students in the areas of computer science and engineering, providing two years of paid tuition and fees and a $42,000 annual stipend for living and conference expenses. Facebook also covers the costs of attending the annual Fellowship Summit at Facebook headquarters.

Zhang was one of 26 fellows selected from a group of 71 finalists who remained from the more than 2,100 applications received. Among the finalists were ICS computer science Ph.D. students Vikram Narayanan (in the area of distributed systems) and Avinash Kumar (in the area of database systems). Zhang was recognized for his work in the area of AR/VR computer graphics.

“Zhang has been working on physics-based differentiable rendering — an emerging area with the potential to greatly benefit applications in many areas including computer graphics, vision, virtual/augmented reality (VR/AR), and computational imaging,” says Zhang’s adviser, Assistant Professor of Computer Science Shuang Zhao. “His recent works have established the mathematical and computational foundations for many problems in this area.”

Zhang, who hopes to go into industry after graduation, says his long-term goal is to “provide an efficient and robust solution to the inverse rendering problem centered around the inference of geometric and material properties from physical measurements (such as from photographs or a depth sensor).”

He first became interested in the topic after earning his bachelor’s degree in electrical engineering. “I have been a big fan of animations and movies, and that is the initial reason why I decided to study computer graphics.” He went on to earn his master’s degree in computer science at Columbia University and then started working with Zhao here at UCI to generate photo-realistic images using computers. 

“For decades, researchers have been working on this and great progress has been achieved, but it is still difficult to create high-quality 3D models to be rendered,” he explains. He notes that even an experienced artist usually needs a few hours to create a high-fidelity 3D model using specialized software such as Maya or Houdini. “I am interested in making this modeling process easier so that everyone can reconstruct a high-quality 3D model maybe just by taking some photos and letting the computer do the rest.”

— Shani Murray