Israel Invents Sarcasm Detector (Yea, We’re really impressed)
A team of Israeli researchers from Hebrew University have created SASI, or Semi-supervised Algorithm for Sarcasm Identification. Said to be 77 per cent accurate, its algorithms were derived from scanning thousands of user reviews from Amazon.com. Oh, so SASI can read. Great.
The whole thing is text based, which unfortunately is pretty useless for breaking down a blogger’s sarcastic wit. Apparently researchers went through 66,000 user reviews from only the brightest minds. Because obviously it takes a great deal of intellect to throw down a coherent thought or two on Amazon.com. I mean, you have to log in etc., it’s a whole to do. The team would flag phrases for sarcasm, which must have been quite exciting, especially after the 100th review of Robert Ludlum’s The Bourne Objective, by Eric Van Lustbader (we’re assuming the ghost writer has a ghost writer). With Israel’s brightest minds on the task, they soon started to see patterns in the customer reviews. It must have been tough, an Israeli who understands sarcasm?
With certain statements being put into a classification system via algorithms, it became pretty much a matter of dotting the I’s and crossing the T’s to get SASI up and running. Now that computers can tell when we’re being sarcastic 77% of the time, it’s only a matter of time until humans catch up to that level. Either that, or SASI will become SkyNet and the machines will rise up against us.
We expect a report from SASI about this post in my office by Monday.
/sarcasm
To read more, click here.
Like
Leave a reply