Multi-Layer Backpropagation Library v1.0


Advertisements

If by any chance you wanted to create any application that needs the power of a neural network, then I have a solution for you. I introduce you the Multi-Layer Feed-Forward Library using the back-propagation training algorithm.

This library for sure will make easier the implementation of a neural network in your application. For now, the library only supports feed-forward networks.

You probably know what a feed-forward network is, if not I’ll explain it briefly. A feed-forward neural network is a structure where the inputs are propagated and processed thought the neurons from the input layer to the output layer. This is far one of the simplest and useful neural networks around.

What you can do with this library

This library allows you to create, train and implement neural networks in your application, and everything by just writing a few lines of code.

Supported programming Languages

So far the library is supported for C++ and Windows, and it will available for Linux and mobile devices soon.

Features

  • Creation of and training neural networks of unlimited sizes and layers.
  • An easy way to store the neural network data via files with a customized ID so your application could only access it.
  • Easy to use data type structures for safely handling lists and floating point arrays
  • Multithreaded execution during training
  • Source code examples
  • Full documentation
  • Full support via email

Requirements

  • Visual C++ compiler or GNU C++
  • Microsoft Visual C++ 2008 runtimes for 32 bits version, Microsoft Visual C++ 201 runtime for 64 bit version or MinGW runtimes depending on which compiler you have.
  • Available on dynamic link library

Price

The Multi-Layer feed forward library is free.

You can use accepting this software is provided ‘as-is’, without any express or implied warranty. In no event will the authors be held liable for any damages arising from the use of this software.
You can use it on non commercial and commercial applications under the following restrictions:

  • You must not take owner ship or say that you wrote this library.
  • An acknowledgment in you application documentation is required along with a link to NeuroAI website URL: http://www.learnartificialneuralnetworks.com
  • You may not reverse engineer, decompile, or disassemble any of the binary files included in this package.
  • You may not modify any source files and libraries included in this package and republish them as yours.
  • Copyright and license notices on source files may not be removed or altered.

Download

UPDATE

Download Most Recent Version MLBP version 1.0.1a.
FIXED SOME BUGS!

Older version libraries here. Feel free to report any bug if you find one.

Documentation: Multi-Layer Backpropagation Library Reference

Examples: Simple Implementation Source code

Leave a Reply

Your email address will not be published. Required fields are marked *