Computer learns meaning of words with Civilization II and wins

Machine-learning system achieves 79 percent rate of victory with strategy game’s manual

TODO alt text

My word processor can tell me if I%26rsquo;ve misspelled a word. Computers handle language as data all the time %26ndash; and do it quite well. But although my copy of Microsoft Word can recognize and highlight something like %26ldquo;Schwarzeneger%26rdquo; with a red squiggly line, it doesn%26rsquo;t understand the meaning of that word or what it signifies %26ndash; and it certainly won%26rsquo;t comprehend complete sentences, instructions, or any memorable one-liners about how %26ldquo;it%26rsquo;s not a tumor.%26rdquo;

But recent research at MIT%26rsquo;s Computer Science and Artificial Intelligence Lab has demonstrated that computers can analyze words, derive meaning from those words, and follow a set of instructions to perform a task. In this case, a special machine-learning system was able to achieve a 79 percent rate of victory in Civilization II against the game%26rsquo;s AI opponents using instructions from the manual.

%26ldquo;Games are used as a test bed for artificial-intelligence techniques simply because of their complexity,%26rdquo; explains S.R.K. Branavan, author of research paper %26ldquo;Learning to Win by Reading Manuals in a Monte-Carolo Framework,%26rdquo; to MIT News. %26ldquo;Every action that you take in the game doesn%26rsquo;t have a predetermined outcome, because the game or the opponent can randomly react to what you do. So you need a technique that can handle very complex scenarios that react in potentially random ways.%26rdquo;

The impressive feat isn%26rsquo;t that the MIT machine-learning system was able to beat Civilization II%26rsquo;s in-game AI; the big deal is that it was able to do it from scratch by interpreting and learning the meaning of words from what associate professor Regina Barzilay describes as the %26ldquo;very open text%26rdquo; in videogame instruction manuals. %26ldquo;They don%26rsquo;t tell you how to win. They just give you very general advice and suggestions, and you have to figure out a lot of other things on your own,%26rdquo; explains Barzilay.

You can read Branavan%26rsquo;s paper, which was presented at this year%26rsquo;s Association for Computational Linguistics (ACL) meeting, here.

July 14, 2011

Source:MIT News

The impressive feat isn%26rsquo;t that the MIT machine-learning system was able to beat Civilization II%26rsquo;s in-game AI; the big deal is that it was able to do it from scratch by interpreting and learning the meaning of words from what associate professor Regina Barzilay describes as the %26ldquo;very open text%26rdquo; in videogame instruction manuals. %26ldquo;They don%26rsquo;t tell you how to win. They just give you very general advice and suggestions, and you have to figure out a lot of other things on your own,%26rdquo; explains Barzilay.

You can read Branavan%26rsquo;s paper, which was presented at this year%26rsquo;s Association for Computational Linguistics (ACL) meeting, here.

July 14, 2011

Source:MIT News

We recommend