to downgrade dimensions ok some algorithm, but opposite, would need to keep extra metadata (ex, 3d > 2d like computer 3d graphics get flatten to 2d screen is easy (x,y), but to expand 2d back to 3d would need that eliminated z to be stored on each point. so for higher dimensional space would it be fair saying metadata extra per point = top dimensional - flatten dimensional (like 3d to 2d screen = 1 "the z (imagining x,y as screen coordinate".
So from 50th dimensional to 3d means each 3d point needed to keep its origin 47 upper axis values? and need of proper compression and elegant algorithm to make it perform while reducing storage.
2025-06-30 00:46:59
0
To see more videos from user @datamlistic, please go to the Tikwm
homepage.