Artificial intelligence (AI) applications are becoming more and more prevalent in our everyday lives. Most contemporary implementations of AI use digital logic and the conventional CMOS hardware that has enabled the information revolution. A team of NIST researchers seeks to enable future generations of AI by focusing on fabricating and measuring new brain-inspired circuits and architectures based on novel devices to deliver computing methods, speeds, and energies better than those achievable using the current computing paradigm. Conventional computing represents information with binary encoding – ones and zeros times different powers of two. The variety of approaches studied at NIST are founded on the concept that computing can be more efficient when information is directly represented by the physical properties of devices themselves. The devices then perform the computation directly, as opposed to the conventional approach of manipulating the binary representation.
The need for these novel approaches is driven by several realities of modern computing. We are pushing computers to take on tasks that humans are much better at than traditional computers. The demands for this type of computing is growing much faster than the capabilities of traditional computers. Perhaps most of all, the energy required to deliver the computations is the most rapidly increasing sector of energy consumption in the world, and it must be reduced to limit the impact on the climate. It is also essential to make computing more efficient in “edge” applications in which computers are embedded in devices that have very restricted energy supplies. The efficiency of the brain drives research that identifies devices acting like the neurons and synapses of the brain and uses them to enable algorithms that compute like the brain. NIST’s AI Hardware team’s research aims to develop the necessary device-level and circuit-level measurements and theory to support the evolution of this technology from laboratory research to commercial application.
The Physical Measurement Laboratory has several projects in Hardware for AI. Click on the links below for more information.
Superconducting hardware for high-speed bio-inspired computing. Bio-inspired computing architectures based on electronic neurons and synapses hold promise for new, faster, more efficient forms of computing. We are developing superconducting versions of these devices that operate at much greater speed and even lower energy than the brain, as well as the metrology needed to understand the interactions that occur in these new ultrafast spiking systems.
Superconducting Optoelectronic Systems. Communication and computation are the twin pillars of neural systems. This project explores the hypothesis that few-photon communication combined with superconducting-electronic computation may enable neuromorphic supercomputers.
Spintronics-based Neuromorphic Computing. Spintronic devices, particularly magnetic tunnel junctions, are compatible with existing CMOS circuitry but have complementary capabilities, such as nonvolatility, tunable dynamics, and nonlinear coupling. We are pursuing a variety of approaches to use these capabilities to develop more efficient computing schemes.
Integrated CMOS Testbeds. To evaluate whether novel devices, such as memristors or magnetic tunnel junctions, are useful for computing, it is necessary to integrate them into CMOS circuitry, which is prohibitively expensive for individual investigators. We are developing for our own use and for dissemination small scale circuitry that can be used to test materials, embed devices in circuitry, and to evaluate moderate scale neuromorphic architectures.
Temporal Computing. Rather than using the values of voltages to represent information, temporal computing using the timing of changes in voltage. We aim to implement such schemes using novel hardware embedded in CMOS circuitry.
Measurements of Memristive Neuromorphic Devices. The physical mechanisms behind the desired behavior of memristors is not well understood. We are developing a suite of measurements to clarify the underlying mechanisms.
The Spin Electronics Group in the Quantum Electromagnetics Division
The Quantum Nanophotonics Group in the Applied Physics Division
The Alternative Computing Group in the Nanoscale Device Characterization Division