Bioinspired neuromorphic module based on carbon nanotube/C60/polymer composite

Abstract

It is extremely challenging to imitate neural networks with their high-speed parallel signal processing, low power consumption, and intelligent learning capability. In this work, we report a spike neuromorphic module composed of ''synapstors'' made from carbon nanotube/C60/polyimide composite and ''CMOS Somas'' made from complementary metal-oxide semiconductor electronic circuits. The ''synapstor'' emulates a biological synapse with spike signal processing, plasticity, and memory; the ''CMOS Soma'' emulates a Soma in a biological neuron with analog parallel signal processing and spike generation. Spikes, short potential pulses, and input to the synapstors trigger postsynaptic currents and generate output spikes from the CMOS Somas in a parallel manner with low power consumption. The module can be modified dynamically on the basis of the synapstor plasticity. Spike neuromorphic modules could potentially be scaled up to emulate biologic neural networks and their functions.

Graphene Dynamic Synapse with Modulatable Plasticity

“…And the plasticity is the foundation of memory and learning behaviors. Because the nervous system based on synapses is a natural paradigm of high performance computation, lots of efforts have been made to achieve an artificial synapse with electronics − and even some artificial neural networks. , …”

“…Because the nervous system based on synapses is a natural paradigm of high performance computation, lots of efforts have been made to achieve an artificial synapse with electronics 5−13 and even some artificial neural networks. 14,15 However, only the static plasticity is not adequate to achieve the complex behaviors of a living organism. With the rapid progress of imaging technology, the dynamic movement in biological synapse has been discovered and has aroused widespread interest.…”

Neuromorphic Computing between Reality and Future Needs

“…The carbon nanotube is another material under consideration for various neuromorphic applications. A variety of neuromorphic components, including dendrites on neurons [143][144][145][146][147], synapses [148][149][150][151][152][153][154][155][156][157][158][159][160][161][162][163][164][165], and spiking neurons [166][167][168][169], have been proposed to use carbon nanotubes. Carbon nanotubes have been used because they can provide the scale (number of neurons and synapses) and density (in terms of synapses) of neuromorphic systems that may be necessary for replicating or imitating biological neural systems.…”

Recent progress on fabrication of memristor and transistor-based neuromorphic devices for high signal processing speed with low power consumption

“…Kim et al continued to report a spike neuromorphic module composed of synapstors and CMOS Soma made from CNT=C60=polyimide (PI) composite and CMOS electronic circuit. 96) They fabricated a CNT network by spin coating a 98% semiconducting single-walled CNT solution onto a SiO 2 surface on a Si substrate. A CNT channel of 10 mm length and 8 mm width was defined by photolithography and connected by Ti=Au (5 nm=100 nm) source and drain electrodes.…”