home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.org.mensa
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!moe.ksu.ksu.edu!usenet-feed.cc.umr.edu!UMRVMA.umr.edu!S106495
- From: S106495@UMRVMA.umr.edu
- Subject: Intelligence etc. (was athiesm and intelligence)
- Date: Thu, 21 Jan 1993 22:51:29 GMT
- Nntp-Posting-Host: umrvma.umr.edu
- Organization: Univ of MO - Rolla
- Sender: cnews@umr.edu (UMR Usenet News Post)
- Message-ID: <1993Jan21.230105.11619@umr.edu>
- Lines: 20
-
- I never meant to say that Neural nets were a substitute for thinking. I
- meant that neural nets and other nonconventional forms of "problem solving" can
- create good answers under certain conditions. These conditions could be
- focusing a camera, or more standard problems that people solve, or even
- problems that we haven't invented yet. Without fully understanding the problem
- , (and thus knowing the answer) there is no way to say any given method will
- work. It is a matter of limits in resources, usually, when deciding how many
- methods to use to solve a problem. In most everyday circumstances, we don't
- have the resources to use anything but habit, because the consequences are
- small.
- Neural networks, GA's etc are probably only the "primitive" beginnings
- of computer softwares and AI's that we won't be able to judge fully until they
- actually exist. Maybe they will never be as good as humans, maybe they will be
- a thousand times as powerful as any human mind.
-
- What is important about humans, I hope, is not simply their ability to use
- logic and do science.
- -Mike B.
- PS: if a computer did act like a human mind, would it make the human mind
- obsolete or just prove that imitation is the sincerest form of flattery?
-