Guest Editorial for Special Issue on Scalable Applications of Neural Networks to Robotics

By Patrick van der Smagt and Daniel BullockThe goal of this special issue is to present a set of papers focused on scalable applications ofneural networks to robotics. The field of neural networks is very broad, with a biological sideand an engineering side. On the biological side, the field spans from studies of highly specializedneuronal circuits in primitive animals to studies of the kind of general-purpose, adaptive neuronalcircuits that are exemplified by the human cerebral cortex and the cerebellum. On theengineering side, the field spans from adaptive function approximation, through machine visionto pattern recognition and classification, and on to adaptive control systems for processes of allkinds. Robotics is an area in which the full range of neural networks ideas could find a naturalhome, but for many applications, adaptive, self-learning methods will become practical andpreferred solutions only if they are implemented in a way that scales well with the dimensionalityof the problem.