Temporally learning floating-gate VLSI synapses

We present a floating-gate synaptic circuit that updates its weight according to the spike-timing-dependent plasticity (STDP) rule. The weight (or floating-gate voltage) is updated only if the time difference between the pre- and post-synaptic spikes falls within a learning window. The update is implemented through tunneling and injection mechanisms which can be tuned for very long time constants up to seconds. The novelty of this circuit is that the tunneling and injection mechanisms are turned on only when the correlation of the pre and postsynaptic activity is significant. The additional benefit of this non-volatile technology is that synaptic weights can be stored locally on chip. We present experimental results that show the learning and normalization effects from the fabricated circuits.

[1]  Paul Hasler.  Foundations of learning in analog VLSI , 1997 .

[2]  Giacomo Indiveri,et al.  Neuromorphic Bistable VLSI Synapses with Spike-Timing-Dependent Plasticity , 2002, NIPS.

[3]  Alan F. Murray,et al.  Synchrony Detection by Analogue VLSI Neurons with Bimodal STDP Synapses , 2003, NIPS.

[4]  Paul E. Hasler,et al.  Supervised learning in a two-input analog floating-gate node , 2004, 2004 IEEE International Symposium on Circuits and Systems (IEEE Cat. No.04CH37512).

[5]  Carver A. Mead,et al.  A single-transistor silicon synapse , 1996 .

[6]  Giacomo Indiveri,et al.  A VLSI spike-driven dynamic synapse which learns only when necessary , 2006, 2006 IEEE International Symposium on Circuits and Systems.

[7]  David Hsu,et al.  Learning Spike-Based Correlations and Conditional Probabilities in Silicon , 2001, NIPS.

[8]  G. Bi,et al.  Synaptic Modifications in Cultured Hippocampal Neurons: Dependence on Spike Timing, Synaptic Strength, and Postsynaptic Cell Type , 1998, The Journal of Neuroscience.

[9]  Miguel Figueroa,et al.  On-Chip Compensation of Device-Mismatch Effects in Analog VLSI Neural Networks , 2004, NIPS.

[10]  Miguel Figueroa,et al.  Adaptive CMOS: from biological inspiration to systems-on-a-chip , 2002, Proc. IEEE.

[11]  Giacomo Indiveri.  Neuromorphic Bisable VLSI Synapses with Spike-Timing-Dependent Plasticity , 2002, NIPS 2002.

[12]  C. Diorio,et al.  A simulation model for floating-gate MOS synapse transistors , 2002, 2002 IEEE International Symposium on Circuits and Systems. Proceedings (Cat. No.02CH37353).

[13]  Shih-Chii Liu,et al.  Silicon synaptic adaptation mechanisms for homeostasis and contrast gain control , 2002, IEEE Trans. Neural Networks.

[14]  J.V. Arthur,et al.  Recurrently connected silicon neurons with active dendrites for one-shot learning , 2004 .

[15]  Paul Hasler,et al.  Continuous-time feedback in floating-gate MOS circuits , 1998, ISCAS '98. Proceedings of the 1998 IEEE International Symposium on Circuits and Systems (Cat. No.98CH36187).

[16]  S. Nelson,et al.  Hebb and homeostasis in neuronal plasticity , 2000, Current Opinion in Neurobiology.

[17]  L. Abbott,et al.  Competitive Hebbian learning through spike-timing-dependent synaptic plasticity , 2000, Nature Neuroscience.

[18]  D. Johnston,et al.  Regulation of Synaptic Efficacy by Coincidence of Postsynaptic APs and EPSPs , 1997 .

[19]  R. Kempter,et al.  Hebbian learning and spiking neurons , 1999 .

[20]  Kwabena Boahen,et al.  Learning in Silicon: Timing is Everything , 2005, NIPS.

[21]  Ph. Häfliger,et al.  Spike Based Normalizing Hebbian Learning in an Analog VLSI Artificial Neuron , 1999, NIPS 1999.