I wrote **** On Jan 20, 2008 2:34 PM, Jim Bromer <[EMAIL PROTECTED]> wrote: > I am disappointed because the question of how a polynomial time solution of > logical satisfiability might affect agi is very important to me.
Well, feel free to start a new thread on that topic, then ;-) In fact, I will do so: I will post a message on SAT, SMT and AGI **** And here it is: However, I would rephrase the question as: How would a pragmatically useful polynomial time solution of logical satisfiability affect AGI? In fact, it's interesting to talk about how existing SAT and SMT solvers http://en.wikipedia.org/wiki/Satisfiability_modulo_theories -- which are often quite effective on surprisingly large real-world problems, in spite of being exponential time in the worst case -- affect AGI. SMT in particular seems to have deep potential applicability. It would seem to me that a practical SMT solver handling quantifier logic would be a more useful research goal than proving P=NP. In AGI, we don't care that much about worst-case complexity, nor even necessarily about average-case complexity for very large N. We care mainly about average-case complexity for realistic N and for the specific probability distribution of problem-cases confronted in embodied experience. Most work with SMT solvers seems to have to do with theories like arithmetic... simple stuff. But what if the theory involved is the (large) set of predicates probabilistically held to be true by an AGI system. How effective are current SMT solvers then? If they are effective, then SMT could prove an interesting tool within an AGI inference engine... a way of relatively rapidly resolving complex queries... -- Ben -- Ben Goertzel, PhD CEO, Novamente LLC and Biomind LLC Director of Research, SIAI [EMAIL PROTECTED] "We are on the edge of change comparable to the rise of human life on Earth." -- Vernor Vinge ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=8660244&id_secret=88001594-8a8210
