Contenu principal
Neuromorphic Computing
One of the other key goals of the HBP is to turn knowledge of the brain into new information technology. This will be the mission of a new Facility for Neuromorphic Computing. As the HBP builds ever more sophisticated models of the brain, the facility will use simplified versions of these models as the basis for increasingly sophisticated hardware devices and systems. Key features are likely to include low energy consumption; resilience and robustness; new techniques of data storage and transmission; extracting information from noisy, error prone, and approximate data; adaptive problem solving; self-repair. In parallel with this work, computational neuroscientists will develop and test cognitive architectures using top down approaches based on mathematical theories of brain function. At a relatively early stage in the project, combinations of these two approaches will lead to new sensors and measuring devices and to novel systems for pattern recognition and categorization, information retrieval and data analysis. As the HBP progresses, it will investigate more complex applications, in which brain-like architectures open the road to systems with genuine intelligence. This prospect promises to enormously enhance and ultimately to transform current models of computing.