Evolution of Artificial Neural Development

In search of learning genes

verfasst von: Gul Muhammad Khan

Verlag: Springer International Publishing

Buchreihe : Studies in Computational Intelligence

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book presents recent research on the evolution of artificial neural development, and searches for learning genes. It is fascinating to see how all biological cells share virtually the same traits, but humans have a decided edge over other species when it comes to intelligence. Although DNA decides the form each particular species takes, does it also account for intelligent behaviour in living beings?

The authors explore the factors that are perceived as intelligent behaviour in living beings and the incorporation of these factors in machines using genetic programming, which ultimately provides a platform for exploring the possibility of machines that can learn by themselves, i.e. that can “learn how to learn”.

The book will be of interest not only to the specialized scientific community pursuing machine intelligence, but also general readers who would like to know more about the incorporation of intelligent behaviour in machines, inspired by the human brain.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Making the Computer ‘Brained’

Abstract

In his play Prometheus Bound; Aeschylus, a Greek dramatist describes how a titan, Prometheus freed mankind of ignorance and enlightened them with the gift of knowledge by deceiving the Greek god Zeus; and how this transgression incurred the wrath of Zeus. This brings in the light significance of knowledge and the fact that since time immemorial, there has been a struggle to acquire knowledge.

Gul Muhammad Khan

Chapter 2. The Biology of Brain: An Insight into the Human Brain

Abstract

Human brain is a very complex but fascinating subject. The structure of brain has always been an attractive topic to the human beings. This small mass miraculously controls every action of the human body. Unfortunately it is not possible to cover every single aspect of the human brain in just one book. This book will try to explain various aspects of brain which are thought to be the main cause of the learning ability of the brain. Brain itself is made up of small building blocks called the “neurons” that are a major source of attraction on their own. The complexity of neurons, its structure and its functionality is a debated topic. Neurons are spread throughout the body and are present in different shapes, however their basic mechanism does not change (Kuffler et al. 1984). Figure 1 shows various types of neurons that exist in the human nervous system. This is why; we are going to begin with a detailed insight into the Human Nervous system.

Gul Muhammad Khan

Chapter 3. Evolutionary Computation

Abstract

This chapter is divided into four main sections.

Evolutionary Computation (EC)
Cartesian Genetic Programming (CGP)
Co-Evolutionary Computation (CC)
Developmental Systems (DS)

Gul Muhammad Khan

Chapter 4. Artificial Neural Network (ANNs)

Abstract

This chapter presents a review of the major forms of the Artificial Neural Networks (ANNs) (Sordo 2002). The particular topic of discussion of this chapter is how learning takes place in these models. Different ways of training the networks are examined.

Gul Muhammad Khan

Chapter 5. Structure and Operation of Cartesian Genetic Programming Developmental Network (CGPDN) Model

Abstract

This chapter presents detailed description of the CGPDN model.

Gul Muhammad Khan

Chapter 6. Wumpus World

Abstract

Having gone through all the design procedures, it is now time to test the learning capabilities of the CGPDN in the Wumpus World environment. Wumpus world is a learning problem scenario which is based on agents and artificial environment.

Gul Muhammad Khan

Chapter 7. Checkers

Abstract

The learning capabilities of CGPDN are explored to ‘recognize’ and ‘learn to play’ the arcade board game. The game of Checkers is selected as the arcade game since it was reconnoitred previously for learning by a number of AI algorithms. Like Wumpus world, checkers is also a grid based game however it is much more challenging and complicated compared to Wumpus world. Checkers is of great importance in the history of Artificial Intelligence and can be used as a test bed for evaluating the learning techniques (Dimand and Dimand 1996). The game of checkers is used here for demonstrating the capability of evolved networks to improve their ability to learn (level of play) by continuously playing against better opponents.

Gul Muhammad Khan

Chapter 8. Concluding Remarks and Future Directions

Abstract

Developing an intelligent network capable of learning and adapting in a complex environment without human intervention has been the main focus of this work. We will discuss to what extent the goals of this research have been accomplished and how significant are its contributions to the field of artificial intelligence.

Gul Muhammad Khan

Backmatter

Titel: Evolution of Artificial Neural Development
verfasst von: Gul Muhammad Khan
Verlag: Springer International Publishing
Electronic ISBN: 978-3-319-67466-7
Print ISBN: 978-3-319-67464-3
DOI: https://doi.org/10.1007/978-3-319-67466-7