The Optimization of Weight Functions of R3,1 Space Time by the Echo State Network

We obtain the optimum weight function of phonetic soliton paths in Weyl fermion sea described by quaternions sitting on (3+1) D lattices. As in the previous work [5] using the Elman Recurrent Neural Network (ERNN) the action of solitonic phonons are derived from the fixed point action prepared for the quantum Chromodynamics (QCD) by DeGrand et al.[20]. We consider 7 paths in (3+1) D space with length less than or equal to 8 lattice units. Comparison of ERNN and the Echo State Network (ESN) are given. The ESN using tanh function for the nonlinear function give stable results as the simulation cycle number increases.