Guanxiong Luo Jason

I am an independent researcher working at the intersection of diffusion modeling, generative AI, and inverse problems in imaging and signal reconstruction. My research focuses on developing self-diffusion and autoregressive diffusion architectures that can recover high-fidelity images from degraded observations while providing principled uncertainty estimation. My CV is available here.

Bridging deep generative modeling, Bayesian inference, and optimization, my projects translate cutting-edge theoretical advances into practical, deployable systems. Using frameworks such as PyTorch, TensorFlow, and TensorRT, I design and implement end-to-end generative pipelines that make state-of-the-art reconstruction and inference methods both efficient and scalable for real-world applications.

Apart from my research, I love soccer, tennis, and photography.