Author Topic: A Landmark in Artificial Intelligence  (Read 1335 times)

0 Members and 1 Guest are viewing this topic.

Offline The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Reef Tank Owner
  • *****
  • Posts: 8115
Re: A Landmark in Artificial Intelligence
« Reply #15 on: December 11, 2017, 10:02:56 AM »
You are both correct.  This is an example of artificially learning to do something intelligently, so it is definitely artificial intelligence.  However, in pop culture, the term artificial intelligence sometimes implies artificial consciousness, which this definitely is not.

Yeah, I think Daniel is at least talking about artificial general intelligence.  This is more limited than that, but it is learning and therefore is intelligence.  In this case, it is an intelligence that is capable of learning to play pretty much any two-person games with clearly defined, unambiguous rules.
« Last Edit: December 11, 2017, 10:07:35 AM by The Latinist »
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline lobsterbash

  • Keeps Priorities Straight
  • ***
  • Posts: 488
Re: A Landmark in Artificial Intelligence
« Reply #16 on: December 11, 2017, 10:05:31 AM »
You are both correct.  This is an example of artificially learning to do something intelligently, so it is definitely artificial intelligence.  However, in pop culture, the term artificial intelligence sometimes implies artificial consciousness, which this definitely is not.

Yeah, I think Daniel is at least talking about artificial general intelligence.  This is more limited than that, but it is learning and therefore is intelligence.  In this case, it is an intelligence that is capable of learning to play pretty much any two-person games with clearly defined, unambiguous rules.

Which is why we shouldn't be impressed by a new AI breakthrough at this point until one can be shown to beat humans at Calvinball.

Offline The Latinist

  • Cyber Greasemonkey
  • Technical Administrator
  • Reef Tank Owner
  • *****
  • Posts: 8115
Re: A Landmark in Artificial Intelligence
« Reply #17 on: December 11, 2017, 10:07:41 AM »
The people behind Stockfish don't like the test conditions (though they give all due credit to the feat).  In particular 1) AlphaZero was running on Google's supercomputer while Stockfish was running on a standard PC and 2) timing was a non-standard 1-minute per move, negating the significant work that has gone into Stockfish's time management; it is designed to conserve time for deeper analysis when it counts.  So not exactly a fair test.
I would like to propose...that...it is undesirable to believe in a proposition when there is no ground whatever for supposing it true. — Bertrand Russell

Offline daniel1948

  • Happy Man in a Boat
  • Reef Tank Owner
  • *********
  • Posts: 9308
  • I'd rather be paddling
Re: A Landmark in Artificial Intelligence
« Reply #18 on: December 11, 2017, 01:02:22 PM »
The people behind Stockfish don't like the test conditions (though they give all due credit to the feat).  In particular 1) AlphaZero was running on Google's supercomputer while Stockfish was running on a standard PC and 2) timing was a non-standard 1-minute per move, negating the significant work that has gone into Stockfish's time management; it is designed to conserve time for deeper analysis when it counts.  So not exactly a fair test.

Thanks for pointing this out. Not a fair test at all with a supercomputer playing against a PC. And tournament chess is always played with clocks that allow you to budget your time.
Daniel
----------------
"Anyone who has ever looked into the glazed eyes of a soldier dying on the battlefield will think long and hard before starting a war."
-- Otto von Bismarck

 

personate-rain
personate-rain