Friday, June 7, 2019

Understanding Unsupervised Learning


AI playing and beating humans in games and what they may do in the future to test them is the topic of this episode of The AI Minute. For more on Artificial Intelligence: https://voicesinai.com https://gigaom.com https://byronreese.com https://amzn.to/2vgENbn... Transcript: In 1949, computer pioneer, Claude Shannon wrote a paper on how a computer may be able to play the game of chess. In the intervening decades we've of course seen computers beat the best humans at a number of games; chess, followed by Jeopardy then by go, and most recently by no limit poker. It raises the question of what might be the next game that the AI's take us on in. Jeff Dean, a senior fellow at Google, has suggested that it may be the game Starcraft. A game where the computer couldn't even effectively evaluate all of the moves available to it let alone multiple moves in advance. Others suggest it may be crossword puzzles by their notoriously cryptic clues. But it could be that there just isn't another game. Certainly not one that will quite capture the imagination of people the way that chess or Jeopardy did. Computers may be beyond all of that at this point. We're at a point where artificial intelligence will make different kinds of milestones than the ones we comment on now. Learning how to tell original stories or judging what's in a video or determining if something's funny or, the granddaddy of them all, mastering unsupervised learning.

No comments: