Extending score based generative modeling to the Riemannian setting.
Paper: https://arxiv.org/abs/2202.02763
Cite
````` @misc{https://doi.org/10.48550/arxiv.2202.02763, doi = {10.48550/ARXIV.2202.02763}, url = {https://arxiv.org/abs/2202.02763}, author = {De Bortoli, Valentin and Mathieu, Emile and Hutchinson, Michael and Thornton, James and Teh, Yee Whye and Doucet, Arnaud}, keywords = {Machine Learning (cs.LG), Probability (math.PR), Machine Learning (stat.ML), FOS: Computer and information sciences, FOS: Computer and information sciences, FOS: Mathematics, FOS: Mathematics}, title = {Riemannian Score-Based Generative Modeling}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Zero v1.0 Universal} } ````` 