Optimization of a nonlinear 1D feature, shown as contour lines, for a 2D non-Gaussian distribution. Inset: Principal component analysis allows extraction of only a linear feature.

From the article:

Renormalized Mutual Information for Artificial Scientific Discovery
Leopoldo Sarra, Andrea Aiello, and Florian Marquardt
Phys. Rev. Lett. 126, 200601 (2021)

View Issue Table of Contents

More Cover Images from Volume 126

Sign In to Your Journals Account

Filter

Filter

Article Lookup

Enter a citation