Artificial Intelligence, Neural Network, Second Life, Applets (AI), Beta Online Books, Artificial Life, Front Page Articles, Graphics, Hardware, Java Programming, Natural Language Processing, Online Books, Popular Articles, Bots, C# Programming, YouTube, Newsroom, Lessons, Outdated Material, Encog, Polls

On Tuesday, June 2, 2009 I will do a presentation on Encog at the Gateway Java Users Group in St. Louis, MO at 6:00 PM. The information on this presentation can be found here. I will post the material for this presentation online after I complete it.

Implementing a Java Neural Network with the Encog Framework

This article shows the very basic process of setting up the Encog Neural Network and Bot framework in the Eclipse IDE. Encog is an advanced neural network and bot programming library. Encog can be used independently either to create neural networks or HTTP bot programs. Encog also includes classes that combine these two advanced features. Encog contains classes for Feedforward Neural Networks, Hopfield Neural Networks, and self organizing maps.

Encog allows your Java and DotNet applications to create and use neural networks. These neural networks can be saved to Encog data files. These files have the .eg extension. Starting in version 1.1 the Encog workbench is now available. The Encog workbench allows you to create and manipulate these EG files. These files can be saved for later use inside of your Java or DotNet applications.

Seen below is a simple Encog datafile. This shows a neural network and training data.

We just released version 1.1 of Encog for DotNet. This also includes a Workbench that allowed editing of Encog datafiles. This is a major update to the Encog Neural Network and Bot framework. v1.1 adds the Encog workbench, which is a GUI tool that can be used to edit the .EG files that Encog uses to save neural networks and training data. The workbench can be used to construct all neural network types that Encog supports, such as Hopfield, Feedforward, and Self Organizing Maps. Training can also be performed using the workbench.

C# and Java are fairly similar languages. They are similar enough that projects, such as our Encog project, can be converted to C#. Yet they are different enough that this is not always a 100% straight forward process. Additionally, there are several considerations to take into account so that your Java program does not look like a "Java Program converted to C#". C# programs support unique indexing options, properties and many other features that are not available to Java programs. For a true translation, it is important to use these as well.

An artificial chemistry is a computer model used to simulate various types of systems. This can be particularly useful for an artificial life simulation. This series of articles will take you through the process as we create a simple artificial chemistry system. The goal is to use this artificial chemistry system to simulate basic life.

Encog 1.0 for Java has been released. Encog 1.0 for C# will be released soon, and is currently in the works. For now, Encog 0.5 is the current C# release. The following is a list of features that were implemented in Encog 1.0.

• All Neural networks should implement the Network interface
• Package structure reorg
• Allow the three neural networks to be persisted in an XML format
• Unify the three neural network types into a single solution
• Add an image based data source
• Add a CSV based data source
  

Introduction to Neural Networks with Java, Second Edition is now available for purchase
E-Book form! You can also download all examples from this book. We will be posting about
half of it online soon. The book will go off to the printer on Monday, and will show up
on Amazon(and the others) in paperback form within a few weeks. The C# neural network
book is nearly complete. It should come out sometime in October, 2008.