Thursday, December 31, 2009

Brains, Meaning and Corpus Statistics


Tom Mitchell, who is head of the machine learning department at Carnegie Mellon, discusses how the human brain represents ideas. His ongoing research is trying to answer questions such as...

- Are brain representations similar across people?
- Can we actually predict brain representations?

His research involves regression analysis of data collected from fMRI scans and word co-occurance statistics using Google's trillion-token corpus.

http://www.youtube.com/watch?v=QbTf2nE3Lbw

No comments:

Post a Comment