Inicio  /  Applied Sciences  /  Vol: 12 Par: 18 (2022)  /  Artículo
ARTÍCULO
TITULO

Fractional Derivative Gradient-Based Optimizers for Neural Networks and Human Activity Recognition

Oscar Herrera-Alcántara    

Resumen

In this paper, fractional calculus principles are considered to implement fractional derivative gradient optimizers for the Tensorflow backend. The performance of these fractional derivative optimizers is compared with that of other well-known ones. Our experiments consider some human activity recognition (HAR) datasets, and the results show that there is a subtle difference between the performance of the proposed method and other existing ones. The main conclusion is that fractional derivative gradient descent optimizers could help to improve the performance of training and validation tasks and opens the possibility to include more fractional calculus concepts to neural networks applied to HAR.

 Artículos similares

       
 
Angel Golev, Snezhana Hristova and Asen Rahnev    
In this paper an algorithm for approximate solving of a boundary value problem for a nonlinear differential equation with a special type of fractional derivative is suggested. This derivative is called a generalized proportional Caputo fractional derivat... ver más
Revista: Algorithms

 
Farhad Amiri, Mohsen Eskandari and Mohammad Hassan Moradi    
The penetration of intermittent wind turbines in power systems imposes challenges to frequency stability. In this light, a new control method is presented in this paper by proposing a modified fractional order proportional integral derivative (FOPID) con... ver más
Revista: Algorithms

 
Rfaat Moner Soliby and Siti Suhana Jamaian    
This paper has two main objectives. First, we modify the traffic flow model by introducing the uphill dispersion that derives from the fact that, in peak hours, drivers tend to travel from low to high density regions. This means that the proposed model r... ver más
Revista: Infrastructures

 
Muhammad Amir Fawwaz, Kishore Bingi, Rosdiazli Ibrahim, P. Arun Mozhi Devan and B. Rajanarayan Prusty    
Managing industrial processes in real-time is challenging due to the nonlinearity and sensitivity of these processes. This unpredictability can cause delays in the regulation of these processes. The PID controller family is commonly used in these situati... ver más
Revista: Algorithms

 
Francesca Pitolli, Chiara Sorgentone and Enza Pellegrino    
Differential problems with the Riesz derivative in space are widely used to model anomalous diffusion. Although the Riesz?Caputo derivative is more suitable for modeling real phenomena, there are few examples in literature where numerical methods are use... ver más
Revista: Algorithms