Re:On Chomsky and the Two Cultures of Statistical Learning
by kenjackson on Friday May 27, 2011 @01:58PM
(#2592418)
Hmm... I never thought of it that way. That sports are a weighted random number generator,but the various weights are unknown. And the commentators are discussing theories as to what the weights are, and how derived. (Although the cartoon seems to be saying the narratives are just about the numbers generated, which is more cynical, and frankly less interesting).
[ Parent
| on HN ]
Re:On Chomsky and the Two Cultures of Statistical Learning
by yourcelf on Friday May 27, 2011 @08:26PM
(#2593555)
Actually, Larry Birnbaum over at Northwestern is doing exactly that: http://infolab.northwestern.edu/projects/stats-monkey/
They take the coded sports results, and automatically generate narratives using statistical speech models. They have a startup that is doing it too, don't recall the name of it....
EDIT: I believe this is it: http://narrativescience.com/