Representation as a Fluent: An AI Challenge for the Next Half Century

Abstract

The authors argued that AI systems must be able to manipulate their own internal representations automatically to deal with an infinitely complex and ever-changing world and to scale up to rich, complex applications. Such manipulation must go beyond changing beliefs and learning new concepts in terms of old concepts; it must be able to change ontology's underlying syntax and semantics. Initial progress has been made, but further progress is urgently needed owing to the demands of autonomous multiagent systems. Understanding and implementing this ability must be a major focus of AI for the next 50 years