Chairman's Report 2007 on
AAAI Quantum Interaction Symposium at Stanford USA in March/April
The AAAI Spring Symposium 2007
This was the first time, that the (just renamed US) Association for the Advancement of Artificial Intelligence, had held a Symposium on quantum models of artificial intelligence, reflecting the increasingly prevalent viewpoint expressed by leading AAAI practitioners and literature, that Turing/digital models of AI might be unsound. Indeed the question of what theoretical models of AI might be sound was central to quantum interaction session debate, which also featured viewpoints centred on other probabilistic and non deterministic alternatives even going so far as to propose that quantum mechanics might itself be replaced by such theory in relation to physics.
This doubt arises because quantum mechanical prediction, although never yet invalidated experimentally, is thought to be incompatible with general relativity used successfully in many explanations of cosmological phenomena, and because string theory as an alternative to quantum mechanics would in principle resolve this dichotomy. However string theory has over many years failed to register any experimental support, and now seems unlikely for sound theoretical reasons, to be able to do so. String theory is in the BCSCMsG view see below in Pauli's famous words “not even wrong” ie it is syntactically but not semantically correct.
The paper presented by Rowlands and myself, 'How Intelligence Evolved?' is based on the discovery of Rowlands and Diaz of the nilpotent universal quantum computational rewrite system (NUCRS) with its universal grammar, as now set out in much detail in Rowlands World Scientific book 'Zero to Infinity' see http://www.worldscibooks.com/physics/6544.html The discovery/book offers a novel semantic viewpoint able to resolve the doubts referred to above in relation to quantum mechanics, general relativity and string theory.
The AAAI paper summarizes Chapter 20 entitled 'Nature's Rules' which references extensive published work over the last twenty years. Well reviewed for inclusion in the Quantum Interaction symposium, the paper outlines the staircase of evolutionary 'molecular' complexity corresponding to the universal infinite alphabet of the NUCRS universal grammar. This starts from an empty state as nilpotence requires, and then progresses (as the result of a fundamental breaking of symmetry) from 3+1 relativistic space time and the matter of Standard Model elementary particle physics (as experimentally validated), through the chemistry of the periodic table of the atomic elements, to that of the RNA/DNA genetic code.
This code (see the book Chapter 19, Nature's Code) is a further instantiation/rewrite of the NUCRS at a higher level of chemical/molecular complexity. That is to say, the genetic code which, as is now generally accepted, includes the coding for the human brain, is a yet further extension of the NUCRS universal alphabet. This would therefore permit the human brain-as a semantic machine- by means of NUCRS rewrite mechanisms to input/output natural language i.e. to hear and to speak and to process language grammatically i.e. to proofread it/think. QED.
The NUCRS, semantically correct because of its universal grammar, thus offers more than Turing computation can by solely assuring syntactically correct programming language. And as it is known that Turing computation is a proven subset of quantum computation, this explains the distinction between the two. This viewpoint that quantum computation offers more not less than Turing/digital computation, in the form of a universal grammar, was also independently confirmed at the AAAI meeting by Bob Coecke of Oxford University in his paper 'Automated Quantum Reasoning: Non-Logic -> Semi-logic -> Hyper Logic. A full record of all the papers presented are to be found in Technical Report SS-07-08 Quantum Interaction, AAAI Press.
