Jingyi Wan

I am a student in Graphics, Vision & Imaging Science track at the University of Cambridge, funded by the Google DeepMind Scholarship. As part of my studies, I work on generative modelling, specifically pose-guided human image synthesis, under the supervision of Prof. Cengiz Öztireli. I'm also fortunate to be mentored by Prof. Marcus A. Brubaker, whose support has deepened my vision for research and my future path.

Previously, I completed my undergraduate studies at Queen Mary University of London, where I worked with Prof. Marcus Pearce on multimodal models for predicting depression levels. I have worked across a range of topics, including NeRF, SLAM, LLMs, and vision-language models, gaining hands-on experience with both foundational concepts and practical implementations.

Email / Github / Google Scholar / LinkedIn

Research

I'm interested in 3D reconstruction, generative AI, and multimodal learning.

Gaussian Head & Shoulders: High Fidelity Neural Upper Body Avatars with Anchor Gaussian Guided Texture Warping
Tianhao Wu, Jing Yang, Zhilin Guo, Jingyi Wan, Fangcheng Zhong, Cengiz Öztireli,
ICLR, 2025
project page / arXiv

Miscellanea

	GPT series and Large Language Models This summary explores the evolution of transformer architectures, explaining core components like attention mechanisms and tracing the development from task-specific models to GPT-1 through GPT-4.
	Generative Models This blog provides an overview of generative models, covering key types like autoregressive models, VAEs, and GANs, along with their learning mechanisms, strengths, limitations, and evaluation methods.
	Diffusion Models This blog provides an a comprehensive overview of diffusion models.

Design and source code from Jon Barron's website.