« Going Green at Conference | Main | The information age is a cyberpunk dystopia »

Google's Peter Norvig on "Throwing more data at the problem"

I attended one talk during the 2008 IST Graduate Symposium this year and it was the opening keynote from Google's Dr. Peter Norvig, director of research. I was hoping to hear more about the new Google Research project that would support eScience and large data sets. Instead I got a talk that appealed to the Computer Scientist in my (ah, 211). He covered what was in the description of the keynote, the amazing ability of simple algorithms to identify images and natural language text... just because of BILLIONS of data points. Test their ability to draw semantic links between ideas from words with Google Sets.

Some gems of insight:
"Don't make algorithms that work well, make algorithms that work well with large enough data"
We can model the entire world by "using the world as it's own model"
We don't need to define rules because "the rules are in the data"

TrackBack

TrackBack URL for this entry:
https://blogs.psu.edu/mt-unprotected/mt-tb.cgi/9368

Comments (1)

David Weinberger makes a lot of the same points in defense of tagging in Everything is Miscellaneous; they may look noisy at first, but the signal gets much stronger as more tags are added because the malicious or irrelevant tags are drowned out by the good ones.

Post a comment

(If you haven't left a comment here before, you may need to be approved by the site owner before your comment will appear. Until then, it won't appear on the entry. Thanks for waiting.)

About

This page contains a single entry from the blog posted on February 7, 2008 9:40 AM.

The previous post in this blog was Going Green at Conference.

The next post in this blog is The information age is a cyberpunk dystopia.

Many more can be found on the main index page or by looking through the archives.

Powered by
Movable Type 3.33