Abstract: The sigmoid function is a representative activation function in shallow neural networks. Its hardware realization is challenging due to the complex exponential and reciprocal operations.

Conventional approaches to the titration of serum antibody binding use mid-point or end-point titers that are in a relative space and are therefore difficult to standardize. Here we propose the use of ...

Abstract: This article addresses the variable convergence rate stability problem for nonlinear impulsive stochastic systems (NISSs). To solve the issue, a novel methodology of sliding mode surface ...