Description du projet
Un réseau neuronal artificiel innovant susceptible d’apprendre 10 millions de fois plus rapidement que nous
Développer des réseaux neuronaux artificiels capables d’apprendre, de traiter l’information et de «penser» de la même façon que le cerveau constitue une sorte de Graal aux applications virtuellement infinies. Les paradigmes d’apprentissage profond exploitant les réseaux neuronaux artificiels hiérarchiques multi-couches qui imitent la structure du cerveau peuvent également imiter sa capacité d’apprentissage, par exemple. Ils ont réalisé des avancées incroyables, dépassant même les performances humaines dans certains cas. Le projet RadioSpin, financé par l’UE, entend faire la démonstration de réseaux d’apprentissage profond traitant des signaux de radiofréquences (RF) et apprenant à des vitesses près de 10 millions de fois supérieures au cerveau humain. Les applications de benchmarking cibleront la mammographie et les empreintes digitales par RF d’IdO.
Objectif
The goal of RadioSpin is to build a hardware neural network that computes using neural dynamics as in the brain, has a deep layered architecture as in the neocortex, but runs and learns faster, by seven orders of magnitude. For this purpose, we will use ultrafast radio-frequency (RF) oscillators to imitate the rich, reconfigurable dynamics of biological neurons. Within the RadioSpin project, we will develop a new breed of nanosynapses, based on spintronics technology, that directly process the RF signals sent by neurons and interconnects them layer-wise. We will demonstrate and benchmark our concept by building a lab-scale prototype that co-integrates for the first time CMOS RF neurons with spintronic RF synapses. We will develop brain-inspired algorithms harnessing oscillations, synchrony and edge-of-chaos for computing and show that they can run on RadioSpin deep network RF technology. Finally, we will benchmark RadioSpin technology for biomedical and RF fingerprinting applications where fast and low energy consumption classification of RF signals are key.
To achieve its ambitious goals RadioSpin brings together frontier researchers along the entire chain of neuromorphic engineering, from material science (spintronic nanodevices), physics (non-linear dynamics), electronics (RF CMOS design), computer science (artificial intelligence algorithms), and microwave signal processing. Two innovative companies bring real-life use-cases (microwave mammography and IoT RF fingerprinting). The scientific experts are further complemented by experts in the field of innovation, commercial deployment and IP monetisation, as well as communication and public engagement.
Champ scientifique
- engineering and technologyelectrical engineering, electronic engineering, information engineeringinformation engineeringtelecommunicationsradio technologyradio frequency
- natural sciencescomputer and information sciencesinternetinternet of things
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringsignal processing
- natural sciencesphysical scienceselectromagnetism and electronicsspintronics
- natural sciencescomputer and information sciencesartificial intelligencecomputational intelligence
Programme(s)
Régime de financement
RIA - Research and Innovation actionCoordinateur
33000 Bordeaux
France