Dimitrios Gerontitis, Activation functions in ZNN neurodynamics

Alexey Vakhnin, Large-scale Global optimization using Evolutionary Algorithms: Where we stand and what's next?

Activation functions in ZNN neurodynamics

Designing new categories of activation functions (AFs) and the comparison of the convergence results between of them is an important field of study due to its utility in the acceleration of the convergence speed in recurrent neural network (RNN) design. In the context of this tutorial, the zeroing neural network (ZNN) based on different nonlinear activation functions are presented for the solution of time-varying Sylvester equation (TVSE). Numerical example in MATLAB software illustrates the efficiency of each AF in the ZNN dynamics.

Large-scale Global optimization using Evolutionary Algorithms: Where we stand and what's next?

Large-scale global optimization (LSGO) problems are a challenging task for a wide range of optimization approaches. The modern world places demand on the development of methods that can effectively explore and find optimal solutions in spaces with a huge number of dimensions. The main challenges and difficulties faced by traditional optimization methods when working with high-dimensional spaces will be discussed. In particular, the impact of the curse of dimensionality on classical methods and optimization approaches will be examined. The tutorial will offer an overview of modern approaches to solving LSGO problems in high-dimensional spaces, including metaheuristic methods, adaptive strategies, and the latest technologies that allow considering the features of such spaces. The purpose is to provide an overview and understanding of which optimization methods may be most effective and promising in working with large-scale spaces, as well as to identify directions for future research and development in this area.