Two major open source neural network projects for the Java platform, Encog and Neuroph, have announced collaboration on the development of advanced Java neural network technology. The basic idea for collaboration is to provide the best of both worlds: Encog’s speed (support for multicore and GPU) and Neuroph’s (a NetBeans Platform application) easy to use neuron-based interface on top for advanced neural network research. This will be accomplished by creating a high performance Encog kernel which will also be used by Neuroph.
The projects will stay separate but will share technologies, and aim for interoperability between their software in future developments.
The most important things Neuroph will gain are: Multicore Support, GPU Support, and some high speed training algorithms (Batch Training, Resilient Propagation Training, and Manhattan Update Training). On the other hand, Encog neural networks will get compatibility with a friendly, rich OO API from Neuroph, which is easy to use and understand.
The projects will benefit from bringing together their communities, and may be able to propose some standards (formats and interfaces) that other similar projects will follow too.
The collaboration will start with a pilot project of creating the Encog Kernel and integrating it with Neuroph. Upon succesful completion of the pilot project, collaboration will continue through many other aspects that have been identified.
- Jeff Heaton is an author, consultant, artificial intelligence (AI) researcher and former college instructor. Heaton has penned more than a dozen books on topics including AI, virtual worlds, spiders and bots. Heaton leads the Encog project, an open source initiative to provide an advanced neural network and bot framework for Java and C#. A Sun Certified Java Programmer and a Senior Member of the IEEE, he holds a Masters Degree in Information Management from Washington University in St. Louis. Heaton lives in St. Louis, Missouri, USA.
- Zoran Sevarac is a teaching assistant at Belgrade University, Department for Software Engineering, researcher at Laboratory for Artificial Intelligence at same University, and member of GOAI Research Network. Through his research, he has been working on development of Java neural network framework which was released as open source project Neuroph. He holds Masters Degree in Information Systems, and he is a PhD candidate. He lives in Belgrade, Serbia, Europe.
- Benchmarking and Comparing Encog, Neuroph and JOONE Neural Networks
- Comparing Neural Networks in Neuroph, Encog and JOONE
- Improving Neuroph Performance