NOTE! This site uses cookies and similar technologies.

If you not change browser settings, you agree to it. Learn more

I understand

Learn more about cookies at :

July 2, 2020 | BA-Healthcare starts production of its CLEAR-M ventilator monitoring system, designed by List

ClearMOn March 13, at the height of the Covid-19 pandemic, CEA Tech institute List initiated the CLEAR project to develop a solution to the shortage of ventilators.

CLEAR (CEA List Emergency Assistance for Respiration) produced results in early April, in the form of the CLEAR-M ventilator monitoring system designed to enhance emergency and transport ventilators. The high-performance, affordable CLEAR-M prototype was tested in the ventilator weaning ward at the Raymond Poincaré Hospital in Garches on Covid-19 patients in recovery but still on ventilators. CLEAR-M was also implemented at the Nord Essonne Hospitals Emergency Department in Orsay.

The performance of the CLEAR-M system was validated, and is garnering growing interest from both manufacturers and hospitals. Based in Rennes, France, BA-Healthcare, a subsidiary of BA-Systèmes, is currently completing the first test manufacturing run of the system, which will equip emergency and transport ventilators so that they can be used to treat Covid-19 patients. “We are proud to be contributing in any way we can to everyone’s efforts to fight Covid-19. Our people are committed to the project, and we are excited to launch full-scale production,” said BA Healthcare General Manager Samuel Pinault.

List has received other expressions of interest from within the Greater Paris hospital network and from multiple manufacturers. CLEAR-M prototypes have been tested at the Raymond Poincaré Hospital in Garches, by Toulouse EMS, and Brest University Medical Center; the early results are encouraging. The system, which will first need to be approved by France’s drug and medical device regulator, could help improve the care given to patients in future epidemics as well as during transport.



June 19, 2020 | Artificial Intelligence of Things (AIoT) proof-of-concept chip presented at VLSI 2020

SamurAI 250
Die micrograph of SamurAI with its building blocks, 4.5mm²

Researchers at CEA Tech institutes Leti and List developed the world’s first low-power IoT node with an integrated artificial intelligence accelerator combined with an ultra-fast wake-up time. They presented the research that led to the groundbreaking chip at the 2020 Symposia on VLSI Technology and Circuits (VLSI) on June 14.

When it comes to artificial intelligence, most computing is done in the cloud, far from the source of the data. Tighter integration—moving computing resources closer to where the data is collected—reduces power consumption, latency, and potential privacy breaches. Highly energy-efficient AI accelerators combined with low-power, versatile IoT nodes would be needed for this kind of integration to be viable.

The CEA’s proof-of-concept SamurAI chip couples a low-power IoT node with an energy-efficient machine learning (ML) accelerator. A dual-subsystem scheme allows this AIoT node to address a wide range of computing applications while delivering optimal energy efficiency.

In low-power mode, an event-driven asynchronous controller core with instant-on capabilities (207ns) executes short, sporadic computing tasks at 1.7MOPS. In high-performance mode, a RISC-V low-power core coupled with an energy-efficient ML accelerator with 64 processing engines executes the most demanding tasks. The node delivers up to 36GOPS and 1.3 TOPS/W when running ML tasks. Together, the dual-subsystem scheme with asynchronous logic and ML accelerator enable a 15,000-fold peak-to-idle power reduction, evidence of the architecture’s versatility.

SamurAI archi 400
SamurAI system architecture, with Always-Responsive and On-Demand sub-systems

The SamurAI chip, made using STMicroelectronics FDSOI28 technology, is equipped with a wake-up radio to receive small messages, a cryptographic accelerator to secure communications, external non-volatile memory for deep sleep mode, and power management with adaptive voltage scaling for further circuit-level energy reduction. It delivers 4 times better computing performance, 3.5 times better power efficiency, and 2 times better power reduction than similar IoT nodes.

It was tested on a people-counting and scene-classification scenario, where it slashed the total power consumption of the system (which included a video camera, sensor, and radio module) by a factor of 3 with the dual-subsystem scheme and by a factor of 2.3 when the ML accelerator is used instead of the RISC-V core. The privacy of the data captured was protected.

Learn more:

SamurAI: A 1.7MOPS-36GOPS Adaptive Versatile IoT Node with 15,000x Peak-to-Idle Power Reduction, 207ns Wake-up Time and 1.3TOPS/W ML Efficiency



June 9, 2020 | Quantum algorithms put to the test

algorithmes quantiques 250List, a CEA Tech institute, recently made an advance in quantum-computer programming with the development of a quantum program specification, programming, and formal verification environment.

While quantum computers are still a while away, one day they will exponentially increase computing capacities in several key fields. And researchers are getting ready now, making sure they will be prepared to verify the performance of the programs that will ultimately be implemented on quantum computers—a task that presents a number of challenges. While there are methods for testing conventional programs, they are not really suitable for quantum computers, either because they would be too costly or, in some cases, simply impossible to implement.

To fill this gap, Carnot CEA List researchers drew on their broad, deep knowledge of computer security to develop a reliable analysis tool. They used state-of-the-art formal methods (techniques to verify that conventional programs do not contain any bugs), but made them compatible with quantum code. First, they made the necessary changes to each logic component (specifications for the behaviors expected of the system, mathematical characterization of the system's actual behavior, and the algorithm to verify that the behavior aligns with specifications). Next, they created QBrick, a specification, programming, and formal verification environment for quantum programs based on the best practices in formal methods for conventional software verification, but modified for quantum computing. They validation-tested this significant advance on a complex use case: the Quantum Phase Estimation (QPE), the core component of Shor's algorithm, which can be used to hack financial transaction encryption keys, a world first!

Read article at



June 4, 2020 | CIVA will soon have its first Structural Health Monitoring tool

SHM civa 250The latest version of the CIVA non-destructive testing (NDT) software platform developed by List, a CEA Tech institute, will be released soon. This latest release has a brand-new structural health monitoring (SHM) module developed specifically for metal and composite structures. This is the first SHM tool to be integrated into CIVA.

Structural health monitoring (SHM) leverages sensors that are permanently integrated into a structure to monitor the structure's condition throughout its lifespan. The latest release of CIVA features a new elastic guided wave monitoring module for metals and composites. CIVA is a non-destructive testing (NDT) software platform developed by List, a member of the Carnot Network.

The new integrated simulation tools show the trajectory of elastic waves in structures with no defects and calculate the disruption to the waves if defects (corrosion, holes, and cracks for metal; delamination for composites) are present in the material. The tools will provide valuable assistance optimizing SHM systems. Designers will now be able to determine how many piezoelectric sensors are needed and where they should be positioned to ensure effective monitoring of a structure.

Another benefit of the new tools, which are available to SHM engineers and experts, is that they automate the implementation of a large number of simulations in just minutes. CIVA also generates very substantial savings on computing costs compared to conventional methods, a crucial factor for large simulation campaigns to feed learning algorithms or statistical analyses.

CIVA will then offer a powerful and innovative SHM solution that will create major new opportunities like certification for SHM systems in industries like aeronautics, aerospace, energy (nuclear, oil & gas), and land transportation. The module will be sold by Extende. Future versions of the software will offer expanded geometry, defect, and sensor libraries and will include advanced data processing capabilities.

Read article at



May 28, 2020 | Towards neural networks that can withstand attacks

At a time when artificial intelligence is making inroads into our everyday lives, List, a CEA Tech institute, is driving advances in cybersecurity that could result in more robust neural networks. CES 2020 provided an opportunity to showcase two demonstrator systems.

From autonomous vehicles to video surveillance, the potential uses for AI in our everyday lives are vast. Hackers, however, are rapidly coming up with attacks on these new applications for AI. Most attacks take advantage of the vulnerability of deep learning systems to disrupt the signal (image, sound) and "trick" or, in some cases influence the AI's decisions. List, a member of the Carnot Network, develops trustworthy AI. The institute recently came up with some effective ways to fend off attacks.

Specifically, they intentionally introduced random modifications of the neural activations from the earliest stages of the neural network design process. The goal is to scramble the network during the learning phase as well as during operation. The researchers' approach enables the machine to remember only the relevant parts of incoming information and to not be fooled by an attack. An alternative for existing machines that need additional protection is to introduce the noise directly into the incoming signal. These modifications are made using an "overlayer" that mitigates or neutralizes the effects of an attack.

The loss of performance that occurs when the defect is introduced is offset by the fact that the system is more robust and can better withstand attacks. A demonstrator presented at CES 2020 was well received. An article* was also published very recently in Neural Information Processing Systems, a major scientific journal in the field of AI.

*Pinot, R., Meunier, L., Araujo, A., Kashima, H., Yger, F., Gouy-Pailler, C., and Atif, J. (2019). Theoretical evidence for adversarial robustness through randomization. In Advances in Neural Information Processing Systems 32, pp. 11838–11848.

Read article at