Add an optimization algorithm to the paper descrription

[thesz]

Original KAN paper used L-BFGS algorithm which is second-order algorithm.

Some of the "compare MLP to KAN" papers here (including "fairer comparison," "Wav-KAN" and "KAN for time series analysis" papers) use Adam optimizer (a first order algorithm) to train both KAN and MLP. This is pretty unfair because Adam is known to perform poorly on the non-neural-network types of functions [1].

[1] https://parameterfree.com/2020/12/06/neural-network-maybe-evolved-to-make-adam-the-best-optimizer/

It should be noted for reader/reseacher convenience that paper uses first-order and/or second-order algorithm(s) for KAN training.