Brown CS News

Brown CS Visual Computing PhD students Yiqing Liang and Rao Fu were recently selected for the annual Rising Stars workshop, a competitive program hosted by the Massachusetts Institute of Technology (MIT)’s Department of Electrical Engineering and Computer Science that recognizes talented underrepresented PhD students and postdocs, especially those who could potentially become faculty members in the coming years.

An advisee of Brown CS faculty member James Tompkin, Yiqing’s research focuses on enabling machines to perceive and reason about the dynamic 3D world over time by unifying sensing, modeling, and reasoning. For sensing, ZeroMSF achieves zero-shot 3D motion estimation across diverse visual domains. For modeling, GauFRe advances dynamic Gaussian Splatting for single-camera 4D reconstruction, while MonoDyGauBench introduces a first benchmark for evaluating dynamic NeRF and Gaussian-based methods. For reasoning, Semantic Attention Flow Fields (SAFF) integrate semantic awareness into radiance fields, and MoDoMoDo improves multimodal large language model reasoning through optimized data mixtures.

For example, her video work ZeroMSF attempts to estimate a 3D point and a 3D vector for every pixel within one video frame that describes where that point will end up in the next video frame. This is a highly ill-posed problem, so Yiqing developed a new large-scale AI approach using synthetic data such that it could apply effectively to real-world scenes. Given this fundamental problem and its wide applicability, Yiqing was awarded a best paper candidate at CVPR (at 0.1% acceptance rate). Other works of hers GauFRe and MonoDyGauBench investigate computationally efficient ways to represent dynamic scenes.

“My studies explore how machines can move beyond static perception to understand and predict the structure and motion of our world,” Yiqing explains. “Ultimately, my goal is to build systems that close the perception–action loop, to build agents that can perceive, predict, and act safely and intelligently.”

Yiqing is a final-year PhD candidate. She earned her Master’s degree from Columbia University, where she worked with faculty members Shuran Song and Shih-Fu Chang, and her Bachelor’s degree in Computer Science from Fudan University. She has completed research internships at NVIDIA Research and Meta Reality Labs. Looking ahead, she plans to couple 4D reconstruction with generative priors for predictive reasoning, infuse large transformers with physical awareness for interactive scene manipulation, and advance toward closed-loop spatiotemporal intelligence. 

“I’m truly honored to be selected as a Rising Star,” Yiqing says. “It’s exciting to connect with others pushing the boundaries of AI and vision, and to imagine what intelligent systems of the future might look like.”

Rao Fu is a final-year Ph.D. candidate working with Brown CS faculty Srinath Sridhar and Daniel Ritchie. During her PhD, she spent a summer at the Max Planck Institute for Informatics, working with Professor Christian Theobalt and Dr. Rishabh Dabral, and gained industry experience at Meta and Autodesk. She earned her Bachelor’s of Engineering from the University of the Chinese Academy of Sciences (UCAS), where she worked with Professors Lin Gao and Xilin Chen, as well as Professor Hao Su (UC San Diego).

Rao’s work focuses on bridging linguistic intelligence with spatial-physical intelligence. “Current AI excels at linguistic abilities but still struggles with spatial and physically demanding tasks,” she says. “My work aims to bridge linguistic intelligence with spatial-physical intelligence. Inspired by cognitive science theories, I translate ‘thoughts of space’ into tangible 3D properties, spaces, and actions. My research enables machines to think and plan in 3D, not just in words. In the future, I’m excited to build a comprehensive mental map of space for machines.” 

She is also “deeply interested in multi-sensory intelligence; beyond vision, there are so many fascinating sensory modalities that shape our understanding of the world”, and notes that “exploring unconditional reflexes is another direction I find intriguing”. Rao adds, “I’m grateful that Brown’s inclusive community has provided me with so much support for my work.”

Speaking about the workshop itself, Rao says, “I’m really excited to participate in the workshop with Yiqing and to connect with other participants from diverse backgrounds. I look forward to learning about their research and peeling back some of the mystery around academia as early-career researchers.”

Yiqing and Rao join Tongyu Zhou, a Brown CS PhD student who was selected for the Rising Stars workshop in 2024 for her research on creative visual authoring systems that bridge 2D and 3D design.

For more information, click the link that follows to contact Brown CS Communications Manager Jesse C. Polhemus.