Backpropagation

honzour

About this app

A simple example of backpropagation algorithm. Learn small neural network basic functions like predefined examples: AND, XOR or 2D distance. Import and export of custom tasks from and to xml or well readable csv. Integration with www browser. Creating new or editing loaded tasks in an editor is also possible. You can show the network anatomy and all weights and also the result with marked training set or exact output for set input. If learning failes, restart whole learning algorithm or just single neuron weights. Alpha learning speed is automatic and configurable. Help on each screen. This program may be useful for students of a basic course of artificial neural networks. It illustrates well that learning of neural networks is a complex task and basic backpropagation without any improvements can only solve very simple tasks.

Ratings and reviews

Has potential but quite limited. Doesn't let you step through the learning by single steps, doesn't show you all the weight values after learning, not enough controls or info given to truly learn things. Also would be more useful if we could import data alone in a csv rather than requiring thing peculiar so-called-csv format that no one else uses.

This app met more than my expectations. Hats off to the dev, for creating such an app, making neural networks so fun to learn. But the only problem, "I FACE" is that if there was a option to get the outputs as float or int instead of double. You may use the Math.round function to do it. And please make some updates to the UI. Google Material Design would be perfect. Thanks to the DEV

Great untapped potential Option to visualize output during training Make creating new training sets more user friendly Options for activation function, training parameters etc.

What’s new

Java export. Adds button. Targeted to 29 SDK.