Friends of Semantic Compositions

January 2009

Sun Mon Tue Wed Thu Fri Sat
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31

Site Statistics

Blog powered by Typepad

« A translation/search engine for Wikipedia | Main | So long, and thanks for all the service calls »

February 22, 2006

Comments

Elihu Vedder

Similarly difficult to build filters like this are quite common in commercial NLP. Some folks I knew at Inso a while back designed a filter to look for inappropriate usernames for a kids chatroom/bulletin board app. Doing the obvious thing, substring matches, leads to both false positives and false negatives as you point out.

Similar ruckuses are caused by vanity license plates, of course. It would be fun to write a finite state recognizer for vile or vulgar vanity license plates, eg. FOR-NIC8 and the like.

Lynn S

Yahoo keeps pissing me off lately. I would boycott them but I really like my Yahoo email account.

The comments to this entry are closed.