A Denoiser Can Do Much More than Just Clean Noise

Researchers are leveraging denoising algorithms to solve other, far more challenging inverse problems.

Optimal Transport for Generative Neural Networks

The theory of optimal transport can solve problems in the unsupervised generation of text, images, and videos.

Inverse problems help to explain why deep learning implementations sometimes become unstable.

The 2020 SIAM Conference on Imaging Science offered snapshots of deep learning’s impact across imaging domains.

Mark Levi introduces a scenario in which a planet's sun sets in the west as normal, then rises in the west once more.

In recent years, exciting work at the interface of data-driven and model-based approaches has blended both paradigms.

The interactions of Bitcoin miners and consequent evolution of wealth inequality comprise a fascinating problem.

The Extreme-scale Scientific Software Development Kit enables collaboration among many teams of developers.

Mark Levi explores common sources of confusion in deriving the Coriolis acceleration on a rotating platform.

Opportunities in network science and data science have grown tremendously, which yields more data and more questions.

Several myths pertaining to mathematical models that attempt to address COVID-19 may limit their usefulness.

Fast forwarding algorithms could allow scientists to overcome limits that stifle real-world quantum simulations.

Tradition regards rational functions as a computational minefield, but recent developments are changing this picture.

An impressive array of theoretical results in optimization have been motivated by machine learning.

Data science is helping to quantify the underrepresentation of women and people of color in art museums.