A computer generated joke: I like my relationships like I like my source, open

What is funny is, in general, hard to define. Humor is a topic that has been endlessly sliced, diced, debated, and analyzed and still we don’t know why one thing is funny (for example, Robin Williams – NSFW) and another isn’t (for example, Gallagher).

Now, in my first sentence I wrote “in general” because there is simple form of verbal humor that – yes, you guessed it – software has been engineered to produce with some degree of success. The branch of humor in question are one liners of the form:

I like my X like I like my Y, Z

For example (and it’s a somewhat cerebral one):

I like my women the way I like my coffee: detrimental to hippocampal neurogenesis, but conducive to short term memory and attentional control. (Attributed to Steven Kaas by tvtropes.com)

I helpfully highlighted the X, Y, and Z portions thereof. For more examples of this humor genre there is a Reddit thread (NSFW) which had a few politically incorrect gems. In contrast, here are three computer-generated examples of this joke:

I like my relationships like I like my source, open

I like my coffee like I like my war, cold

I like my boys like I like my sectors, bad

Those jokes are the output of a program discussed in the paper “Unsupervised joke generation from big data” by Sasa Petrovic´ and David Matthews of University of Edinburgh School of Informatics. The paper’s abstract concludes:

This is, to the best of our knowledge, the first fully unsupervised humor generation system. Our model significantly outperforms a competitive baseline and generates funny jokes 16% of the time, compared to 33% for human-generated jokes.

Quite fascinating stuff and perhaps by the time SkyNet takes over, it might be a lot funnier than it appeared to be in The Terminator. You can just hear it: “I like my humans the way I like my coffee: I don’t.”