Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!tektronix!uw-beaver!cornell!vax135!ariel!hou5f!hou5g!hou5h!eagle!mhuxi!houxm!ihnp4!ixn5c!inuxc!pur-ee!uiucdcs!uiuccsb!loeb
From: loeb@uiuccsb.UUCP
Newsgroups: net.ai
Subject: Re: intelligence is... - (nf)
Message-ID: <2829@uiucdcs.UUCP>
Date: Thu, 15-Sep-83 22:31:41 EDT
Article-I.D.: uiucdcs.2829
Posted: Thu Sep 15 22:31:41 1983
Date-Received: Sat, 17-Sep-83 15:00:32 EDT
Lines: 18

#R:sri-arpa:-483200:uiuccsb:5500001:000:876
uiuccsb!loeb    Sep 15 17:05:00 1983


    Perhaps it would be helpful to consider the Most complicated learner.
  If there is no limit, then, as the mystics say, the perfect learner will
  deduce from any observation the nature of the entire universe.
    From this limit case, it would seem that the complexity of a learner
  is related to its deductive power.  As human society, which contains
  the whole of science, has greater deductive power than an individual,
  it would seem to be a more complex learner.
    Another obvious factor is experience, but it is interesting that induction
  is not necessary in the limit.  This is probably because the perfect learner
  learns everything instantly from a single experience -- ie it deduces all
  ordinarily induced conclusions.  So speed is another factor.

    Given all this nonsense, what does it mean if there Is a limit to
  the complexity of a learner?