industry news
Subscribe Now

CEA-Leti Reports Machine-Learning Breakthrough That Opens Way to Edge Learning

Article in Nature Electronics Details Method that Takes Advantage of RRAM Non-Idealities To Create Intelligent Systems that Have Potential Medical-Diagnostic Applications

GRENOBLE, France – Jan. 19, 2021 – CEA-Leti scientists have demonstrated a machine-learning technique exploiting what have been previously considered as “non-ideal” traits of resistive-RAM (RRAM) devices, overcoming barriers to developing RRAM-based edge-learning systems.

Reported in a paper published in the January issue of Nature Electronics titled, “In-situ learning using intrinsic memristor variability via Markov chain Monte Carlo sampling”, the research team demonstrated how RRAM, or memristor, technology can be used to create intelligent systems that learn locally at the edge, independent of the cloud. The learning algorithms used in current RRAM-based edge approaches cannot be reconciled with device programming randomness, or variability, as well as other intrinsic non-idealities of the technology.

To get around that problem, the team developed a method that actively exploits that memristor randomness, implementing a Markov Chain Monte Carlo (MCMC) sampling learning algorithm in a fabricated chip that acts as a Bayesian machine-learning model.

The article notes that while machine learning provides the enabling models and algorithms for edge-learning systems, increased attention concerning how these algorithms map onto hardware is required to bring machine learning to the edge. Machine-learning models are normally trained using general purpose hardware based on a von Neumann architecture, which is unsuited for edge learning because of the energy required to continuously move information between separated processing and memory centers on-chip.

Intensive research in the microelectronics industry is currently focused on using RRAM as non-volatile analog devices in hardware-based artificial neural networks that can allow computation to be carried out in-memory, to drastically reduce these energy requirements. RRAM has been applied to in-memory implementations of backpropagation algorithms to implement in-situ learning on edge systems. However, because backpropogation requires high-precision memory elements, previous work has largely focused on how RRAM randomness can be mitigated – often necessitating energy-intensive techniques.

The CEA-Leti-based team’s breakthrough was an approach capable of leveraging this randomness, instead of trying to prevent it, and allowing in-situ learning to be realized in a highly efficient fashion through the application of nanosecond voltage pulses to nanoscale memory devices. This culminated in an extremely low-energy solution. The team explained that, relative to a CMOS implementation of its algorithm, the approach requires five orders of magnitude less energy. That is the rough equivalence of the difference in height between the tallest building in the world and a coin lying on the ground.

As a result, this approach is capable of bringing learning to edge-computing systems, which is impossible using existing commercial approaches. Such an application could be an implanted medical system that locally updates its operation based on the evolving state of a patient. To run a representative test of learning at the edge in such an environment, the team experimentally applied RRAM-based MCMC to train a multilayer Bayesian neural network to detect heart arrhythmias from electrocardiogram recordings – reporting a better detection rate than a standard neural network based on a von Neumann computing system.

“This highlights that, beyond being RRAM-compatible, Bayesian machine learning offers an alternative modelling method that appears well suited to the characteristics of edge learning,” the article says.

The team also applied their experimental system to solve further classification tasks including the diagnosis of malignant breast-tissue samples.

“Our system could be used as the foundation for the design and fabrication of a standalone and fully integrated RRAM-based MCMC sampling chip, for applications outside the laboratory,” the authors of the article conclude. That achievement will finally open the door to edge learning and an entirely new set of applications.

About CEA-Leti (France)

Leti, a technology research institute at CEA, is a global leader in miniaturization technologies enabling smart, energy-efficient and secure solutions for industry. Founded in 1967, CEA-Leti pioneers micro-& nanotechnologies, tailoring differentiating applicative solutions for global companies, SMEs and startups. CEA-Leti tackles critical challenges in healthcare, energy and digital migration. From sensors to data processing and computing solutions, CEA-Leti’s multidisciplinary teams deliver solid expertise, leveraging world-class pre-industrialization facilities. With a staff of more than 1,900, a portfolio of 3,100 patents, 10,000 sq. meters of cleanroom space and a clear IP policy, the institute is based in Grenoble, France, and has offices in Silicon Valley and Tokyo. CEA-Leti has launched 65 startups and is a

member of the Carnot Institutes network. Follow us on and @CEA_Leti. 

Technological expertise

CEA has a key role in transferring scientific knowledge and innovation from research to industry. This high-level technological research is carried out in particular in electronic and integrated systems, from microscale to nanoscale. It has a wide range of industrial applications in the fields of transport, health, safety and telecommunications, contributing to the creation of high-quality and competitive products.

featured blogs
Feb 26, 2021
OMG! Three 32-bit processor cores each running at 300 MHz, each with its own floating-point unit (FPU), and each with more memory than you than throw a stick at!...
Feb 26, 2021
In the SPECTRE 20.1 base release, we released Spectre® XDP-HB as part of the new Spectre X-RF simulation technology. Spectre XDP-HB uses a highly distributed multi-machine multi-core simulation... [[ Click on the title to access the full blog on the Cadence Community si...
Feb 25, 2021
Learn how ASIL-certified EDA tools help automotive designers create safe, secure, and reliable Advanced Driver Assistance Systems (ADAS) for smart vehicles. The post Upping the Safety Game Plan for Automotive SoCs appeared first on From Silicon To Software....
Feb 24, 2021
mmWave applications are all the rage. Why? Simply put, the 5G tidal wave is coming. Also, ADAS systems use 24 GHz for SRR applications and 77 GHz for LRR applications. Obviously, the world needs mmWave tech! Traditional mmWave technology spans the 30 – 300 GHz frequency...

featured video

Silicon-Proven Automotive-Grade DesignWare IP

Sponsored by Synopsys

Get the latest on Synopsys' automotive IP portfolio supporting ISO 26262 functional safety, reliability, and quality management standards, with an available architecture for SoC development and safety management.

Click here for more information

featured paper

Functional Safety-Relevant Wireless Communication in Automotive Battery Management Systems

Sponsored by Texas Instruments

With increasing energy density in HEV/EVs, effective battery management and monitoring is essential to avoid any kind of hazards related to overvoltage or overtemperature. This paper explores achieving ASIL D functional safety compliance while using a wireless battery management system.

Click here to download the whitepaper

featured chalk talk

Time Sensitive Networking for Industrial Automation

Sponsored by Mouser Electronics and Intel

In control applications with strict deterministic requirements, such as those found in automotive and industrial domains, Time Sensitive Networking offers a way to send time-critical traffic over a standard Ethernet infrastructure. This enables the convergence of all traffic classes and multiple applications in one network. In this episode of Chalk Talk, Amelia Dalton chats with Josh Levine of Intel and Patrick Loschmidt of TTTech about standards, specifications, and capabilities of time-sensitive networking (TSN).

Click here for more information about Intel Cyclone® V FPGAs