Oy vey, this page is in need of an update! Coming soon(er or later, probably later): I will be posting various code for implementing the simple layered version of the ‘minibrain’ neural network system, invented by Dante Chialvo and Per Bak, that I have been using in some of my research. This will include a much more efficient version of the code found in the file mbcode.zip, along with a few other pieces of code I have been working on and some general functions which should allow you to implement the minibrain neural network in code of your own*.
Some programs may require functions found in Numerical Recipes in C by W. H. Press et al. (Cambridge University Press, 1988).
* A more efficient version of the minibrain code can be found in learnmap-neg-f.zip but it’s still far from being as good as it could be.
- mb-soc-test-im.zip (5kB)
- [ANSI-compatible C] Measures the distribution of learning times for the minibrain when adapting between different arbitrary input-output maps. Used to generate results in Fig. 4 of Order-disorder transition in the Chialvo-Bak ‘minibrain’ controlled by network geometry [J. Wakeling (2003), Physica A 325, 561-569].
- learnmap-neg-f.zip (5kB)
- [ANSI-compatible C] Measures learning time by the ‘minibrain’ network of input-output maps of different size, relative to the size of the intermediary (hidden) layer of neurons. Used to generate results in Figs. 2 and 3 of Order-disorder transition in the Chialvo-Bak ‘minibrain’ controlled by network geometry [J. Wakeling (2003), Physica A 325, 561-569].
- mbcode.zip (19kB)
- [ANSI-compatible C] Implements the Minority Game with agents controlled by ‘minibrain’ neural networks. Used to generate results shown in Intelligent systems in the context of surrounding environment [J. Wakeling & P. Bak (2001), Phys. Rev. E 64, 051920]. Don’t use this code as it is VERY inefficient; it’s included here only for historical interest!