NeuralBF is proof-of-concept compiler for the Brainfuck "programming language" to recurrent artificial neural networks (R-ANNs).

This project is mainly for fun, but it shows the expressive power of R-ANNs (and, in particular, that they are "Turing complete"). I already worked on a similar idea a long time ago, where I constructed R-ANNs out of a more descriptive custom programming language and even analysed on toy examples how well they could be used as a starting point for further optimisation through genetic algorithms.

More details about what exactly NeuralBF can do and how it works can be found in the README file in the source repository.

The project is free software and licensed under the GNU GPL v3, with the source hosted at Gitlab. See there for more information, but you can also directly download the latest source-code release NeuralBF 0.1 (signature) from here if you prefer.

Copyright © 2011–2019 by Daniel KraftHomeContactImprint