Neuromorphic Memristors Run AI Tasks at 1/800th Power

Good news! Very impressive results! Brains made of silicon!

"... Now two new studies may help solve key problems these components face not just with yields and reliability, but with finding applications beyond neural nets. ...
Such brain-inspired neuromorphic hardware may also prove ideal for implementing neural networks—AI systems increasingly finding use in applications such as analyzing medical scans and empowering autonomous vehicles. ...
researchers in Israel and China fabricated memristive devices using a standard CMOS production line. The resulting silicon synapses the team built boasted a 100 percent yield with 170- to 350-fold greater energy efficiency than a high-performance Nvidia Tesla V100 graphics processing unit when it came to multiply-accumulate operations, the most basic operation in neural networks. ...
The scientists developed the new devices using the floating-gate transistor technology used in commercial flash memory. Whereas conventional floating-gate transistors have three terminals, the new components have only two. ...
The new devices displayed high endurance, operating past more than 100,000 cycles of programming and erasing by using voltage pulses. In addition, they showed only moderate device-to-device variation and are projected to possess long data retention times of more than 10 years. ...
In another study, a team of French researchers investigated using memristors for the statistical computing technique known as Bayesian reasoning, in which prior knowledge helps compute the chances that an uncertain choice might be correct. Its results are fully explainable—unlike many nearly inscrutable AI computations—and it can perform well when there is little available data, as it can incorporate prior expert knowledge. ...

The scientists rewrote Bayesian equations so a memristor array could perform statistical analyses that harnesses randomness—a.k.a. “stochastic computing.” Using this approach, the array generated streams of semi-random bits at each tick of the clock. These bits were often zeroes but were sometimes ones. The proportion of zeroes to ones encoded the probabilities needed for the statistical calculations the array performed. This digital strategy uses relatively simple circuitry compared to nonstochastic methods, all of which reduces the system’s size and energy demands.

The researchers fabricated a prototype circuit incorporating 2,048 hafnium oxide memristors on top of 30,080 CMOS transistors on the same chip. In experiments, they had the new circuit recognize a person’s handwritten signature from signals beamed from a device worn on the wrist. ..."

From the abstract (1):

"Memristors, and other emerging memory technologies, can be used to create energy-efficient implementations of neural networks. However, for certain edge applications (in which there is access to limited amounts of data and where explainable decisions are required), neural networks may not provide an acceptable form of intelligence. Bayesian reasoning could resolve these concerns, but it is computationally expensive and—unlike neural networks—does not naturally translate to memristor-based architectures. Here we report a memristor-based Bayesian machine. The architecture of the machine is obtained by writing Bayes’ law in a way that makes its implementation natural by the principles of distributed memory and stochastic computing, allowing the circuit to function solely using local memory and minimal data movement. We fabricate a prototype circuit that incorporates 2,048 memristors and 30,080 transistors using a hybrid complementary metal–oxide–semiconductor/memristor process. We show that a scaled-up design of the machine is more energy efficient in a practical gesture recognition task than a standard implementation of Bayesian inference on a microcontroller unit. Our Bayesian machine also offers instant on/off operation and is resilient to single-event upsets."

From the abstract (2):

"Memristor-based neuromorphic computing could overcome the limitations of traditional von Neumann computing architectures—in which data are shuffled between separate memory and processing units—and improve the performance of deep neural networks. However, this will require accurate synaptic-like device performance, and memristors typically suffer from poor yield and a limited number of reliable conductance states. Here we report floating-gate memristive synaptic devices that are fabricated in a commercial complementary metal–oxide–semiconductor process. These silicon synapses offer analogue tunability, high endurance, long retention time, predictable cycling degradation, moderate device-to-device variation and high yield. They also provide two orders of magnitude higher energy efficiency for multiply–accumulate operations than graphics processing units. We use two 12 × 8 arrays of memristive devices for the in situ training of a 19 × 8 memristive restricted Boltzmann machine for pattern recognition via a gradient descent algorithm based on contrastive divergence. We then create a memristive deep belief neural network consisting of three memristive restricted Boltzmann machines. We test this system using the modified National Institute of Standards and Technology dataset, demonstrating a recognition accuracy of up to 97.05%."

Memristors Run AI Tasks at 1/800th Power - IEEE Spectrum These brain-mimicking devices boast tiny energy budgets and hardened circuits

(1) A memristor-based Bayesian machine (no public access, but article above contains a link to the PDF file)

(1) A memristive deep belief neural network based on silicon synapses (no public access)