CEA-List recently came up with a method for selecting the most suitable existing neural network for adaptation and reuse for new target applications. This advance makes it possible to train classification neural networks, or CNNs, without subject matter expertise or large datasets.
CEA-List has developed a new multimode method to improve the quality controls performed on the conductive inks used to make printed electronics. The goal is to help make sure the finished products operate as intended long after they are manufactured.
CEA-List developed a real-time livestock monitoring system that leverages the analysis of video streams to identify individual animals and specific behaviors that can indicate disease or whether an animal is in heat, for example.
The VMachina 2 project has resulted in the development of a multi-user virtual reality platform for training industrial operators on machines and other job tasks.
CEA-List developed a comprehensive workstation ergonomics assessment tool that analyzes the worker’s posture and effort to rate how difficult a job is. The tool was developed as part of the Ergoforce project at FactoryLab.
Federated learning: CEA-List’s new platform offers a range of innovative tools partners can use to train their models while maintaining a very high level of data security. The platform will be constantly updated with advanced modules to handle heterogeneous data and improve cybersecurity.
Secure cloud solution provider Scille turned to CEA-List for help protecting data in unsecured areas of the cloud. The institute developed a blockchain-based tool to make Scille’s Parsec cloud workspace more robust.
Blockchains will have to be able to communicate with each other if the development of the technology is to pick up speed. Blockchain startup Toposware joined forces with CEA-List to develop a blockchain interoperability protocol that does away with a central authority.