Artificial Creativity and Self-Evolution: Abductive Reasoning in Artificial Life Forms

摘要

What contributions can Cognitive Science offer to the understanding of nature of providing creativity to artificial life forms? For this discussion, it is necessary to investigate creative processes from a mechanistic perspective as well as involve subjective elements which cannot, in principle, be described from this perspective. These two basic approaches will be investigated here, focusing the artificial creative process on the nature of artificial abductive reasoning. As an initial hypothesis we will characterize creativity as a self-organizing process in which abductive reasoning occurs through the use of self-organizing, semantic topical maps in conjunction with an abductive neural network, allowing the creation and expansion of well a structured set of beliefs within the artificial system. This process is considered here as part of the establishment of order parameters in the flow of information available to allow artificial life forms to self-organize and infer on sensory information. In this sense, we will argue that a deeper understanding of how self-organizing processes involving abductive reasoning may take place in artificial dynamic systems, and how this can assist in the creation of an artificial creative process within an artificially intelligent artificial life form we refer to as a Synthetic, Evolving Life Form (SELF) Here we present a self-evolving, abductive, hypothesis-based reasoning framework called the Advanced Learning Abductive Network (ALAN) that provides the ability to mimic human experience-based reasoning.