works as assotiative memory and can:
- 1 Pictures
--- a) Learning
patterns form ASCII format PPM
--- b) Simulating
Hopfield neural network and remeber from part of information all of stored patterns.
- 2 Predictor
--- a) Predicting
the future for example predict orders, weather, sales, traffic jams.
- 3 InteliGUI
--- a) Inteligent GUI
simulation for learn seqences of clicked buttons.
This is Hopfield neural network simulator project. It is writen in C++
programming languages. I use Visual Studio 2008
to develop this project but you can also use SharpDevelop or similar IDE. For parallel computing I use
feature of Visual C++
My goal is to write GUI for this project for watching in real time how Hopfield neural network works.
Picture Learning process:
Picture Simulate process:
This HopfieldSimulator is better than others presented on Internet becouse it can works with
full RGB color scale pictures
not only with black and white pixels. And output vector of this neural network can store any kind of information not only pictures. I decide to show its as picture but it is only mine choice. Values of output vector are
between -1.0 and +1.0 and you can use useful resolution between minimum -1.0 and maximum +1.0 values in your experiments and do anythink you want with that math algorithm.
New implemented feature is Predictor
for guessing future for example the producer can predict purchase orders from shops and plan production. Examples for prediction feature is ofcorse many, many more like weather, traffic jams or problems of machines.
New implemented feature is also InteliGUI
for learn seqences of clicked buttons and to find seqence from begin part of it. You can use this alghoritm to build AI GUI that will be learn seqences of clicked options like menu item, buttons and so on.
I will be develop GUI for this project here and I will be glad to show you my favourite aspects of AI.
For more information in Polish visit my blog:
Best regards and enjoy ;-).
Piotr Sowa, MCAD.NET