NeurIPS 2022 Workshop on Score-Based Methods

New Orleans, December 2nd, 2022

The score function is the gradient of the log-density. As a unique way to represent probability distributions, score functions have enabled the development of many computationally efficient tools for machine learning and statistics, collectively known as score-based methods. By working with distributions through score functions, they have found widespread applications in vastly disjointed subfields of machine learning, such as generative modeling, hypothesis testing, differential equations, MCMC sampling, gradient estimation, control variates, variational inference, inverse problems, denoising, and optimal transport. We encourage submission in any of these areas, as well as novel applications of score-based methods to other fields of science and engineering, such as computer vision, computational chemistry, and astronomy.


Invited Speakers and Panelists

Invited speakers