Pulse-firing neural chips for hundreds of neurons
Advances in neural information processing systems 2
Adaptation of Current Signals with Floating-Gate Circuits
Analog Integrated Circuits and Signal Processing
Neural Networks-Extraordinary Variation
IEEE Micro
Adaptation of Current Signals with Floating-Gate Circuits
MICRONEURO '99 Proceedings of the 7th International Conference on Microelectronics for Neural, Fuzzy and Bio-Inspired Systems
Pulse-Based Circuits and Methods for Probabilistic Neural Computation
MICRONEURO '99 Proceedings of the 7th International Conference on Microelectronics for Neural, Fuzzy and Bio-Inspired Systems
Hi-index | 0.00 |
EPSILON, a large, working, VLSI device, demonstrates pulse stream methods in the wider context of analog neural networks. EPSILON uses dynamic weight storage techniques, but a nonvolatile alternative is desirable. To that end, we have developed an amorphous silicon memory, which we present in experiments incorporating the device in a modest pulse stream neural chip. We have also developed a target-based training algorithm, which we demonstrate in a prototype learning device using a realistic problem. Finally, we explore system-level problems in experiments with a second version of EPSILON in a small, autonomous robot.