Many years ago - 15, I think - with the post "Artificial Stupidity," I wondered
How do you make a computer chess program play not-so-strongly? You can limit the amount of time that it analyzes any position, or you can limit the depth to which it will analyze. Or, you can cause it to blunder every once-in-a-while...
I have recently learned that some great thinking has gone into that question, with fascinating results
Artificial Stupidity: One Google Engineer’s Algorithms for Bad
Chess Playing
Read it and weep. Or chuckle.
No comments:
Post a Comment