Neuroph is lightweight Java Neural Network Framework that can be used to simulate common neural network architectures. With a small number of basic classes that correspond to basic NN concepts, it is easy to learn. It also has a nice GUI application.
|Tags||Scientific/Engineering Artificial Intelligence Neuroscience Software Development Libraries Application Frameworks Java Libraries|
|Operating Systems||OS Independent|
Release Notes: Backpropagation improvements (dynamic learning rate and momentum, min serror stopping criteria, and bias neurons). OCR tools and an API. A stock market prediction sample. Perceptron and Multi-Layer Perceptron learning visualization samples, new methods to make the API more comfortable to use, improved documentation, and a license change from the LGPL 3 to the Apache 2 license.
Release Notes: This release brings important bugfixes and improvements. An issue was fixed with editing a GUI in NetBeans. The LMS formula was fixed. Testing in black and white mode for image recognition was fixed. A GUI bug was fixed regarding exceptions when creating large networks. The image recognition API was changed so the color mode is automaticaly detected from settings used for network training. Graph view was migrated to JUNG to 2.0. Specific network layouts were created and unnecessary options were removed. The ANT build file is now included, which can build the jars for the library and GUI.
Release Notes: The most important feature of this relase is image recognition support. This provides a GUI tool for training multi-layer perceptrons for image recognition and an easy-to-use API to deploy these networks in the end-user applications. With this library, you can perform image recognition in just few lines of code. This release also comes with important API improvments and better documentation.
Release Notes: This release brings momentum for back propagation, importing training sets from txt files, network error graph during the training, support for Instar, Outstar, and BAM networks, XML support for neural networks and training sets, a basic help system, and a sample image recognition application. It also provides a full NetBeans project tree, which will make it easier to develop and build a project according to its original structure.
Release Notes: The core framework library and GUI application were completely separated. Most of the compilation warnings were cleaned up. Deprecated stop() methods for stopping the learning thread were removed. Serialization support was improved. Several bugs were fixed in the core and GUI. Neural network plugins are supported. A labels plugin was added, which enables labeling for all neural network componets. Borland layout managers in the GUI were removed. Documentation was written for all classes and methods from the library. New GUI components were added.