Parametrized, Deformed and General Neural Networks

Více o knize

The book explores innovative activation functions in neural networks, emphasizing parametrized and deformed functions that retain more neurons compared to traditional methods. It highlights the brain's asymmetry through these deformed functions and introduces a wide range of general activation functions. The author's original work covers various neural network types, including ordinary, fractional, fuzzy, and stochastic approximations, as well as univariate and multivariate methods. Additionally, it examines iterated sequential multi-layer approximations within the context of Banach space-valued functions.