On the fusion and transference of knowledge. II

• Main Author: Rubin, S.H.
• Format: Conference Proceeding
• Language: English
• Subjects: Cost function; Humans; Neural networks; Object oriented modeling; Ontologies; Operational amplifiers; Reflection; USA Councils; Vehicles; Working environment noise
• DOI: 10.1109/IRI.2003.1251408

Progress in AI (artificial intelligence) is bottlenecked by contemporary thinking. Soft computing methodologies recognize the fundamental role of imprecision in any computational intelligence. This paper goes one step further in that it recognizes the fundamental need for heuristics - both as a generalization of numeric imprecision according to S.H. Rubin (1999) and, as a consequence, of Godel's Incompleteness Theorem by V.A. Uspenskii (1987). Heuristic ontologies were compared with logical ontologies and shown to be advantageous for the realization of intelligent computation. It was also noted how the mechanics of evolutionary biology and that presupposed for thought can be computationally equivalent. In particular, randomness and symmetry by G.J. Chaitin (1975) are shown to underpin computational evolution by providing for heuristic discovery, knowledge transference and creative computation. The capability for the transference of heuristic knowledge was evidenced to be fundamental to the evolution of what M. Minsky (1987) has termed, "a society of mind". The need for transformational representations of this knowledge was demonstrated by S. Amarel (1968). A modeling language, designed to exploit domain symmetries, was characterized for the game of Tic-Tac-Toe. Applications to UAVs and intelligent agents - among a plethora of others - will follow. In conclusion, Marvin Minsky argues that many AI researchers seem to have lost their way in recent years and should be investigating core problems such as how a person reasons according to C.T. Clark. It is hoped that this paper will sow the seeds for a proper response to Minsky's admonition by inciting a restoration in the type of creative thinking that attracted many AI researchers into the field in the first place.