Ziyu Wan

Shanghai Jiao Tong University, City University of Hong Kong

Papers

Total Citations

H-Index

About

Ziyu Wan is a rising researcher at the forefront of artificial intelligence, making significant strides in both intelligent decision-making and photorealistic 3D generation. His work addresses two critical frontiers: enabling machines to make robust decisions in unpredictable real-world environments, and creating high-fidelity 3D content for applications in AR/VR, robotics, and gaming. In his foundational 2023 paper, Wan proposed a "Foundation Decision Model" perspective, arguing that intelligent decision-making systems must continuously acquire new skills to handle pervasive uncertainty—a paradigm shift from static AI models. This work has garnered early attention with 5 citations, establishing a new research direction. In 2024, Wan co-authored "CAD: Photorealistic 3D Generation via Adversarial Distillation," which tackles the limitations of Score Distillation Sampling (SDS) in 3D synthesis. By introducing adversarial distillation, his method dramatically improves the photorealism and quality of generated 3D objects, addressing the surging demand for 3D data in industry. With 4 citations already, this work promises to accelerate the deployment of generative 3D pipelines. Wan’s dual focus on foundational AI theory and practical 3D generation marks him as a versatile innovator whose contributions are shaping the next generation of intelligent, visually-capable systems.