Computer Learns to Tell Dirty Jokes
May 2, 2011 5 Comments
Finally, a computer capable of passing MY Turing test. Two computer scientists at the University of Washington are creating an algorithm teaching a computer to make “That’s what she said” jokes:
They then evaluated nouns, adjectives and verbs with a “sexiness” function to determine whether a sentence is a potential TWSS. Examples of nouns with a high sexiness function are “rod” and “meat”, while raunchy adjectives are “hot” and “wet”.
Their automated system, known as Double Entendre via Noun Transfer or DEviaNT, rates sentences for their TWSS potential by looking for particular elements such as nouns that can be interpreted in multiple ways. The researchers trained DEviaNT by gathering jokes from twssstories.com and non-TWSS text from sites such as wikiquote.org.
First, the name is outstanding. They report 70% accuracy, but they expect the number to improve with more data to draw on. Of course, 70% success isn’t that bad considering how often my friends make questionable TWSS jokes. (Hint: if you have to explain the context in which she said it, you’ve probably failed.)
In case this isn’t your particular style of humor, they’re hoping it’ll be able to learn new types of humor based on the metaphor mapping.
Suddenly “I can’t let you do that, Dave” has a whole new meaning…
I wonder if this is related to http://www.quora.com/How-would-you-programmatically-parse-a-sentence-and-decide-whether-to-answer-thats-what-she-said …
Are they vying for an Ig Nobel prize?
But seriously, silly applications of serious research make the research sound silly.
Why does research have to be serious? Given that we can hardly tell what basic research will end up being important and wide-spread, why not encourage research into things both silly and serious, mundane and magnificent? After all who’d’ve fricking thought that devices to amplify light by stimulated emission of radiation would be so important, even without putting them on the heads of sharks?
We think of Science as being focused on proper methodology, well-planned and never varying in observation, but the “illogical” silly side is just as important. Without it you turn inventive researchers into drones who who never have that flash of insight that allows them to design and develop the creative experiments that become foundational in their field.
Keep the silly in science!
Because the first thing they do is take the silliness out of it anyway: “We frame the problem of TWSS recognition as a type of metaphor identification.”
Well, I suppose it’s a step up from the computer program that makes “Your face” jokes… 😉