If abc -> abd, then xyz -> ?
December 2, 2004 9:50 AM Subscribe
Copycat [Java] - long available only in its rickety original LISP source - is now an interactive applet. You can propose analogy puzzles based on strings of letters and watch Copycat build up and tear down models of the problem until it finds a solution it's happy with. It will explain its reasoning, and, since stochastic elements are inolved, can come up with alternate solutions upon rerun. Tutorial for the applet here; or, explore MetaCat, the next generation follow-up.
oh that's sweet. thank-you!
if you're interested in this kind of thing, and haven't already read it, grab hold of hofstadter's "fluid concepts and creative analogies".
after several runs it gave me kji -> lji for abc -> abd and i felt like a proud parent ;o)
posted by andrew cooke at 12:09 PM on December 2, 2004
if you're interested in this kind of thing, and haven't already read it, grab hold of hofstadter's "fluid concepts and creative analogies".
after several runs it gave me kji -> lji for abc -> abd and i felt like a proud parent ;o)
posted by andrew cooke at 12:09 PM on December 2, 2004
Ahh, "slipnet" and "coderack" - this must be a Hofstadter project. Excellent, I loved his book, it will be fun to play with this.
posted by Mars Saxman at 12:14 PM on December 2, 2004
posted by Mars Saxman at 12:14 PM on December 2, 2004
Bloody Hell!
As a Hofstadter junkie I got Fluid Concepts & Creative Analogies when it came out.
It's great to see that some of the stuff he and his colleagues wrote about is available online.
Thank you Wolfdog!
posted by thatwhichfalls at 11:06 PM on December 2, 2004
As a Hofstadter junkie I got Fluid Concepts & Creative Analogies when it came out.
It's great to see that some of the stuff he and his colleagues wrote about is available online.
Thank you Wolfdog!
posted by thatwhichfalls at 11:06 PM on December 2, 2004
You know what I've learned from this? I learned how carefully Hofstadter et al must have cherry-picked the examples in the FC&CA book. This program can solve, and reproduce the solution distributions for, the small handful of examples he shows in the book, but not much more. It was, for example, unable to solve
(ccc -> c) => (aaa -> ?)
or,
(abc -> ab) => (pqr -> ?)
or any number of others I thought were truly trivial.
Did anyone get any unexpectedly interesting / insightful results?
posted by Wolfdog at 7:06 AM on December 3, 2004
(ccc -> c) => (aaa -> ?)
or,
(abc -> ab) => (pqr -> ?)
or any number of others I thought were truly trivial.
Did anyone get any unexpectedly interesting / insightful results?
posted by Wolfdog at 7:06 AM on December 3, 2004
« Older Gawking in Atlanta, ya'll | Heroin, PCP, ketamine, marijuana? Newer »
This thread has been archived and is closed to new comments
posted by shoepal at 11:51 AM on December 2, 2004