Switched Capacitor-Based Integrate-And-Fire Neural Network
This article deals with an analog implementation of the functional parts of an Integrate-and-fire neural network. The designed parts would be used as leaf cells with automated layout generation for a complex Integrate-and-fire neural network with built-in analog learning and weight storage. The most important cell is the synapse, which includes a postsynaptic forming block, learning and weight storage, which makes it possible to give individual output characteristics to each synapse in a complete neural network. This approach is conceptually close to its biological counterpart. The design uses a switched capacitor technique to achieve a significant reduction in silicon area.
KeywordsDendritic Tree Clock Signal Switch Capacitor Current Comparator Switch Capacitor Circuit
Unable to display preview. Download preview PDF.
- 1.Zahn T., Izak R., Trott K. (1997): Mixed analog-digital neurochip for acoustical attention. Europian workshop on neuromorphic systems, Stirling, U.K.Google Scholar
- 2.Vittoz E., Oguey H., Maher M. A., Nys O., Dijkstra E., Chevroulet M. (1991): Analog storage of adjustable synaptic weights. In VLSI Design of Neural Networks, Kluwer Acad. Publ., 47–63.Google Scholar
- 3.Glover M. A., Hamilton A., Smith L. S. (1998): Analogue VLSI Integrate and Fire Neural Network for Clustering Onset and Offset Signals in Sound Segmentation System. In Neuromorphic Systems, Engineering Silicon from Neurobiology, Progress in Neural Processing, vol. 10, World Scientific Pub Co, ISBN: 9810233779Google Scholar
- 4.Izak R., Hajtas D., Durackova D. (1998): Analog neuromorphic VLSI system with Integrate and Fire Neurons and on-chip learning Synapses, in Proceedings of 43. IWK, vol. 2, 76–81.Google Scholar
- 5.Ismail M., Fiez T. (1994): Analog VLSI Signal and Information Processing, McGrow - Hill Series in Electrical and Computer EngeneeringGoogle Scholar
- 6.Gregorian R., Ternes G. C. (1986): Analog MOS Integrated Circuits for Signal Processing, John Wiley & Sons, ISBN: 0471097977Google Scholar