Train a Markov chain random sentence generator with the Wikileaks diplomatic cables. The purpose is to generate and transmit more "chaff" to thwart electronic surveillance, e.g., the Emacs spook function.
Use Google Ngrams to determine a Markov model for "regular" English text, then bias the Wikileaks model away from normal text, to accentuate those words and phrases that are especially present in the leaked documents but not anywhere else. (Better would be some corpus on unclassified cables.)
In current events: http://animalnewyork.com/2012/02/the-department-of-homeland-security-is-searching-your-facebook-and-twitter-for-these-words/
No comments :
Post a Comment