zbMATH — the first resource for mathematics

Initial evolution results on CAM-brain machines (CBMs). (English) Zbl 1001.68712
Dorffner, Georg (ed.) et al., Artificial neural networks - ICANN 2001. International conference, Vienna, Austria, August 21-25, 2001. Proceedings. Berlin: Springer. Lect. Notes Comput. Sci. 2130, 814-819 (2001).
Summary: This paper presents results of some of the first evolution experiments undertaken on actual CAM-Brain Machines (CBM), using the hardware itself and not software simulations. A CBM is a specialised piece of programmable (evolvable) hardware that uses Xilinx XC6264 programmable FPGA chips to grow and evolve, at electronic speeds, 3D cellular automata (CA) based neural network circuit modules of some 1000 neurons each. A complete run of a genetic algorithm (e.g. with 100 generations and a population size of 100) is executed in a few seconds. 64000 of these modules can be evolved separately according to the fitness definitions of human “EEs” (evolutionary engineers) and downloaded one by one into a gigabyte of RAM. Human “BAs” (brain architects) then interconnect these modules “by hand” according to their artificial brain architectures. The CBM then updates the binary neural signaling of the artificial brain (with 64000 “hand” interconnected modules, i.e. 75 million neurons) at a rate of 130 billion CA cell updates a second, which is fast enough for real time control of robots. Before such multi-moduled artificial brains can be constructed, it is essential that the quality of the evolution (the “evolvability”) of individual modules be adequate. This paper reports on the first evolution results obtained on CBM hardware.
For the entire collection see [Zbl 0972.68679].
68U99 Computing methodologies and applications
68T05 Learning and adaptive systems in artificial intelligence
92B20 Neural networks for/in biological studies, artificial life and related topics
Full Text: Link