Description du projet
Des réservoirs de nano-aimants pour augmenter notre puissance de calcul
La puissance de calcul disponible peine à répondre à la demande d’analyse car la production de données augmente de façon exponentielle. Pour remédier à ce problème, il est indispensable de développer de nouvelles plateformes de traitement de données massivement parallèles, capables de traiter de grandes quantités d’informations «d’un seul coup» plutôt que morceau par morceau. Pour y parvenir, SpinENGINE propose de combiner deux concepts de pointe: le reservoir computing et la dynamique d’ensemble de nano-aimants (nanomagnet ensemble dynamics). Le reservoir computing utilise un réservoir à la dynamique fortement non linéaire qui projette des signaux d’entrée sur des espaces de haute dimension et utilise des techniques simples de traitement linéaire pour extraire des signaux de sortie. SpinENGINE utilise les interactions non linéaires émergentes et modulables des ensembles de nano-aimants comme réservoir pour créer un nouveau dispositif de calcul massivement parallèle.
Objectif
The SpinENGINE project will lay the foundations for a new, massively parallel, computational platform based on emergent behaviour in large nanomagnet ensembles. The project will develop an efficient, highly scalable, and easily reproducible platform meeting the data analysis challenges in our increasingly data-rich society. We will build upon our recent discoveries and use complex, nonlinear, and highly tunable interactions in such ensembles to realize a hardware platform for “Reservoir Computing”, a biologically-inspired computational approach. Our critical hypothesis is that the synergies between the inherent properties of nanomagnet ensembles and those required for reservoir computing will enable the efficient creation of a highly adaptive computational platform for the analysis of complex, dynamic data sets. This has the potential to greatly outperform current approaches using conventional CMOS hardware.
SpinENGINE will bring together a multidisciplinary team of researchers with expertise in computer science, condensed matter physics, material science, computational modelling, and high-resolution microscopy. This will enable us to simultaneously explore the fundamental behaviours of nanomagnet ensembles and understand how these can be harnessed for useful computation. By the end of the project, we aim to fabricate a proof-of-concept device capable of solving pattern recognition and classification problems, and, in collaboration with our industrial partner, IBM, produce a roadmap to the further scaling and commercialization of our computational platform. Success in the SpinENGINE project will have vast implications for data analysis at all scales, ranging from low power computation in the simplest sensor node to accelerated data processing in the most complex supercomputer.
Champ scientifique
- natural sciencesphysical sciencescondensed matter physics
- natural sciencesphysical sciencesopticsmicroscopy
- natural sciencescomputer and information sciencesartificial intelligencepattern recognition
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectronic engineeringcomputer hardwaresupercomputers
- natural sciencescomputer and information sciencesdata sciencedata processing
Mots‑clés
Programme(s)
Régime de financement
RIA - Research and Innovation actionCoordinateur
7491 Trondheim
Norvège