Bio-inspired perception processor. A neuromorphic approach

Abstract: Innovation in the biomimetic field has become very active since 2015. In France, this is reflected in: the CESE promulgating biomimicry as a matter of public interest, the inauguration of CEEBIOS, the creation by the CNRS of a BioComp working group, etc.

These innovations in neuromorphic processors are now part of the field of artificial intelligence applications. They compete with the deep learning used by major international companies. The contribution of biomimicry to computational processes, presented in this article, is an important differentiator between these techniques.

Abstract: The innovation in the biomimetic domain becoming very active since 2015. It's acted in France by: the CESE by promulgating the biomimicry of public interest, the inauguration of the CEEBIOS, and the creation by the CNRS of a workgroup BioComp, etc.

These innovations done on neuromorphic processors join from now on in the field of the artificial intelligence applications. They compete with the deep learning used by big international companies. Biomimicry contribution in the computation processes, presented in this article, is differential important one between these techniques.

Keywords: bio-inspired processor, semantic representation, spatio-temporal histogram, perception, dynamic attractor, neuromorphic process, perceptual invariance, unsupervised learning

Keywords: neuromorphic processor, bio-inspired, semantic representation, spatiotemporal histogram, perception, dynamic attractor, neuromorphic process, perceptive invariance, unsupervised learning