From Symbolic Computation to Super-Symbolic Computation

EasyChair Preprint 6559

Abstract

Current state of the art information technologies (IT) utilize two forms of computation namely, symbolic computation and sub-symbolic computation. Symbolic computation is performed using algorithms on data structures, which contain knowledge about a particular state of the system in the form of a sequence of symbols. Sub-symbolic computation is performed by neural networks.

Biological systems also use both symbolic and sub-symbolic computations in the form of genes and neural networks. However, biological systems have evolved one step further by incorporating super-symbolic computation that performs computations on the combined knowledge from both symbolic and sub-symbolic computations to derive higher order autopoietic and cognitive behaviors.

The new type of computing automata called a structural machine provides means for modeling symbolic, sub-symbolic and super-symbolic computations performed on data and knowledge structures. In this work we argue that super-symbolic computation adds one more dimension to the general schema of computations. Synthesizing it with symbolic computation and sub-symbolic computation in one model, we come to symbiotic computation. Structural machines with flexible types of processors can accomplish symbiotic computation. Symbiotic computation combines advantages of subsymbolic, symbolic and super-symbolic computations and advance the current state of the art IT with higher-order autopoietic and cognitive behaviors.

Keyphrases: Information Technology, Structural Machines, Symbolic Computing, computational models, information, neural networks

@booklet{EasyChair:6559,
  author    = {Mark Burgin and Rao Mikkilineni},
  title     = {From Symbolic Computation to Super-Symbolic Computation},
  howpublished = {EasyChair Preprint 6559},
  year      = {EasyChair, 2021}}