Synaptic Resistors for Concurrent Inference and Learning with High Energy Efficiency.

The fastest supercomputer, Summit, has a speed comparable to the human brain, but is much less energy-efficient (≈1010 FLOPS W-1 , floating point operations per second per watt) than the brain (≈1015 FLOPS W-1 ). The brain processes and learns from "big data" concurrently via trillions of synapses in parallel analog mode. By contrast, computers execute algorithms on physically separated logic and memory transistors in serial digital mode, which fundamentally restrains computers from handling "big data" efficiently. The existing electronic devices can perform inference with high speeds and energy efficiencies, but they still lack the synaptic functions to facilitate concurrent convolutional inference and correlative learning efficiently like the brain. In this work, synaptic resistors are reported to emulate the analog convolutional signal processing, correlative learning, and nonvolatile memory functions of synapses. By circumventing the fundamental limitations of computers, a synaptic resistor circuit performs speech inference and learning concurrently in parallel analog mode with an energy efficiency of ≈1.6 × 1017 FLOPS W-1 , which is about seven orders of magnitudes higher than that of the Summit supercomputer. Scaled-up synstor circuits could circumvent the fundamental limitations in computers, and facilitate real-time inference and learning from "big data" with high efficiency and speed in intelligent systems.

Doping modulated carbon nanotube synapstors for a spike neuromorphic module.

A doping-modulated carbon nanotube (CNT) electronic device, called a "synapstor," emulates the function of a biological synapse. The CNT synapstor has a field-effect transistor structure with a random CNT network as its channel. An aluminium oxide (Al2 O3 ) film is deposited over half of the CNT channel in the synapstor, converting the covered part of the CNT from p-type to n-type, forming a p-n junction in the CNT channel and increasing the Schottky barrier between the n-type CNT and its metal contact. This scheme significantly improves the postsynaptic current (PSC) from the synapstor, extends the tuning range of the plasticity, and reduces the power consumption of the CNT synapstor. A spike neuromorphic module is fabricated by integrating the CNT synapstors with a Si-based "soma" circuit. Spike parallel processing, memory, and plasticity functions of the module are demonstrated. The module could potentially be integrated and scaled up to emulate a biological neural network with parallel high-speed signal processing, low power consumption, memory, and learning capabilities.

Analog neuromorphic module based on carbon nanotube synapses.

We report an analog neuromorphic module composed of p-type carbon nanotube (CNT) synapses and an integrate-and-fire (I&F) circuit. The CNT synapse has a field-effect transistor structure with a random CNT network as its channel and an aluminum oxide dielectric layer implanted with indium ions as its gate. A positive voltage pulse (spike) applied on the gate attracts electrons into the defect sites of the gate dielectric layer, and the trapped electrons are gradually released after the pulse is removed. The electrons modify the hole concentration and induce a dynamic postsynaptic current in the CNT channel. Multiple input spikes induce excitatory or inhibitory postsynaptic currents via excitatory or inhibitory CNT synapses, which flow toward an I&F circuit to trigger output spikes. The dynamic transfer function between the input and output spikes of the neuromorphic module is analyzed. The module could potentially be scaled up to emulate biological neural networks and their functions.