A neural model for hardware plasticity in artificial vision systems

The advent of massively parallel many-core architectures on a chip can be considered as a good opportunity to rethink the classical computation model used for several decades and that now shows some limitations to follow both the potential and the usage of new technologies. In this paper, the way explored to study new solutions is directly inspired from biology, and more precisely from neurosciences. This way could lead, for example, to best practices for dynamically partitioning application tasks onto a set of processing cores. Distributed load-balancing is known to bring more efficient utilization of resources in the case of regular applications. We propose in this paper a bio-inspired hardware substrate that brings a plasticity property into many-core architectures. We describe a hardware controller in which a grid of processing elements will support a set of neurocognitive processes in order to drive a robot in different tasks. We propose an original distributed hardware artificial neural network as support for this plasticity. It is inspired by Neural Fields equations that have shown self-organizing behaviors and can be suitable for this purpose. It can be used to take allocation decisions locally, taking into account the state of the whole system through the emergent behaviour of the network. This paper describes the neural model and focuses mainly on its implementation onto FPGA in the context of artificial vision.

[1]  A. Treisman,et al.  A feature-integration theory of attention , 1980, Cognitive Psychology.

[2]  Yann Boniface,et al.  Dynamic self-organising map , 2011, Neurocomputing.

[3]  Guang R. Gao,et al.  A Discussion in Favor of Dynamic Scheduling for Regular Applications in Many-core Architectures , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[4]  Nicolas P. Rougier,et al.  Self-Organizing Dynamic Neural Fields , 2013 .

[5]  Gregory D. Peterson,et al.  Evolvable Block-Based Neural Network Design for Applications in Dynamic Environments , 2010, VLSI Design.

[6]  Benoît Miramond,et al.  Using High-Level RTOS Models for HW/SW Embedded Architecture Exploration: Case Study on Mobile Robotic Vision , 2008, EURASIP J. Embed. Syst..

[7]  Paul Bach-y-Rita,et al.  Late human brain plasticity : vestibular substitution with a tongue BrainPort human-machine interface , 2005 .

[8]  Laurent Fiack,et al.  FPGA-based vision perception architecture for robotic missions , 2012 .

[9]  Markus Peura,et al.  The Self-Organizing Map of Trees , 1998, Neural Processing Letters.

[10]  James L. Crowley,et al.  Fast Computation of Scale Normalised Gaussian Receptive Fields , 2003, Scale-Space.

[11]  Olivier Sentieys,et al.  Real-time scheduling on heterogeneous system-on-chip architectures using an optimised artificial neural network , 2011, J. Syst. Archit..

[12]  Bertrand Granado,et al.  Embodied Computing: Self-adaptation in Bio-inspired Reconfigurable Architectures , 2012, 2012 IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum.

[13]  D. Buonomano,et al.  Cortical plasticity: from synapses to maps. , 1998, Annual review of neuroscience.

[14]  Teuvo Kohonen,et al.  Self-organization and associative memory: 3rd edition , 1989 .

[15]  Teuvo Kohonen,et al.  The self-organizing map , 1990 .

[16]  S. Amari.  Dynamics of pattern formation in lateral-inhibition type neural fields , 1977, Biological Cybernetics.

[17]  R. Miikkulainen,et al.  Self-Organization, Plasticity, and Low-Level Visual Phenomena in a Laterally Connected Map Model of the Primary Visual Cortex , 1997 .