Neurobiological Computation and Synthetic Intelligence

Abstract

When considering the ongoing challenges faced by cognitivist approaches to artificial intelligence, differences in perspective emerge when the synthesis of intelligence turns to neurobiology for principles and foundations. Cognitivist approaches to the development of engineered systems having properties of autonomy and intelligence are limited in their lack of grounding and emphasis upon linguistically derived models of the nature of intelligence. The alternative of taking inspiration more directly from biological nervous systems can go far beyond twentieth century models of artificial neural networks (ANNs), which greatly oversimplified brain and neural functions. The synthesis of intelligence based upon biological foundations must draw upon and become part of the ongoing rapid expansion of the science of biological intelligence. This includes an exploration of broader conceptions of information processing, including different modalities of information processing in neural and glial substrates. The medium of designed intelligence must also expand to include biological, organic and inorganic molecular systems capable of realizing asynchronous, analog and self-* architectures that digital computers can only simulate.