| Speaker: | Dr. Aurel Lazar, Columbia University |
|---|---|
| Abstract: | As the critical dimensions in semiconductor technologies continue to scale into the nanoscale regime, the power supply of the devices has to be appropriately reduced to maintain reliable operation. Technology scaling and the associated reduction of the supply voltage leads to a reduction of power consumption for digital ICs while still allowing the clock rate to increase. For analog front-end circuits the power supply scaling is becoming a fundamental limitation, however. These circuits provide the indispensable interface between the analog physical world and the digital signal processing stage. Traditionally analog signals are encoded in the amplitude domain. The reduction of the signal amplitude due to the drastically lower power supplies in nanoscale semiconductors results in a very significant reduction in the accuracy of the analog circuits. A paradigm shift is in order. Instead of encoding the analog signal in the amplitude domain, time domain encoding should be used. Encoding signal information in the time domain leverages the trends in nanoscale semiconductors towards higher operational speeds and thus higher precision for time measurements. Time encoding is a real-time asynchronous mechanism for mapping amplitude information into a time sequence. A Time Encoding Machine (TEM) is a realization of such an encoding mechanism. TEMs represent an alternative to classical samplers; they arise naturally in modeling sensors in communications and sensory systems in neuroscience. A canonical TEM is introduced and two of its instantiations are analyzed. The first TEM is based on the Asynchronous Sigma-Delta Modulator and maps an analog signal into a one-dimensional time sequence. We demonstrate that, under Nyquist type rate conditions, quantization of a bandlimited signal in the time and amplitude domains are largely equivalent methods of information representation. We discuss sensitivity issues of the time decoding algorithm and present an algorithm for perfect signal recovery that is threshold insensitive. The second instantiation is a multichannel TEM based on filter banks and integrate-and-fire neurons that generates a multidimensional time sequence. We provide an algorithm for perfect stimulus (signal) recovery from the multidimensional time sequence. Finally, we show how to construct TEMs based on the Hodgkin-Huxley neuron and give an algorithm for perfect recovery. |
| Biography: | Aurel A. Lazar is a Professor of Electrical Engineering at Columbia University. In the mid 80s and 90s, he pioneered investigations into networking games and programmable networks. In addition, he conducted research in broadband networking with quality of service constraints; and in architectures, network management and control of telecommunications networks. His current research interests are at the intersection of Computational Neuroscience, Information/ Communications Theory and Systems Biology. In silico, his focus is on Time Encoding and Information Representation in Sensory Systems, and, Spike Processing and Neural Computation in the Cortex. In vivo, his focus is on the olfactory system of the Drosophila. |
| Presented On: | Friday, May 6, 2005 |
| Videotape: | <Not available.> |
