"It not only recovered a nonsensical law—it recovered different laws for different galaxies." Consider that there are infinitely more nonlinear possibilities than linear -- how to constrain the training to winnow that down?
Replies
Also suggest applying what you have done to real-world unsolved problems, such as erratic climate indices bsky.app/profile/puki...
jul 16, 2025, 7:16 am • 0 0 • view