Adaptive Programming of Unconventional Nano-Architectures

Novel assembly processes for nanocircuits could present compelling alternatives to the detailed design and placement currently used for computers. The resulting architectures however may not be programmable by standard means. In this paper, nanocomputers with unconventional architectures are programmed using adaptive methods. The internals of the device are treated as a "black box" and programming is achieved by manipulating "control voltages". Learning algorithms are used to set the controls. As examples, logic gates and simple arithmetic circuits are implemented. Additionally, similar methods allow for reconfiguration of the devices, and makes them resistant to certain kinds of faults.

[1]  Kagan Tumer,et al.  An Introduction to Collective Intelligence , 1999, ArXiv.

[2]  Y. Leblebici,et al.  A modular approach for reliable nanoelectronic and very-deep submicron circuit design based on analog neural network principles , 2003, 2003 Third IEEE Conference on Nanotechnology, 2003. IEEE-NANO 2003..

[3]  David H. Wolpert,et al.  Information Theory - The Bridge Connecting Bounded Rational Game Theory and Statistical Physics , 2004, ArXiv.

[4]  David H. Wolpert,et al.  Distributed control by Lagrangian steepest descent , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[5]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.