~ Nonlinear Adaptive Filtering as a Form of Artificial Intelligence ~


Kolmogorov-Arnold model demo UStandard
U-Tree model demo U-Tree
This is a next step in the development of Kolmogorov-Arnold model. The Urysohn blocks of a model are arranged in a form of binary tree. Each non-bottom node has two inputs and one output. Bottom nodes have multiple randomly selected original inputs. The number of bottom nodes is always 2n.

As it was found in multiple experiments U-tree works in the same way of better than Kolmogorov-Arnold or neural network. The first link at top allows to download code along with multiple datasets for Kolmogorov-Arnold concept and second link provides U-Tree demo for the same datasets.