Tuesday, April 29, 2014

Algorithmic Information Theory, Shaken not Stirred with Joshua Porth

In this episode, we serve up an algorithmic information theory cocktail that consists of Shannon's information theory and Turing's computability theory shaken not stirred (to borrow from Gregory Chaitin).

Joining us to share this cocktail is Joshua Porth, as we talk about an interesting and practical application of algorithmic information theory that relies on the Normalized Compression Distance (NCD) to find the similarity between two documents--in this case, between two trouble tickets in an IT department.

 Listen now: (download)

Reference Paper: Clustering by compression
Wikipedia's entry: Normalized Compression Distance
Ray Solomonoff's Papers: http://world.std.com/~rjs/pubs.html

Book Bett mentioned: The Information: A History, a Theory, a Flood

Good background information: Khan Academy's Introduction to Information Theory

No comments:

Post a Comment