The Language Of Humour

Humour is one of the most interesting and puzzling aspects of human
behavior. Despite the attention it has received in fields such as
philosophy, linguistics, and psychology, there have been only few attempts
to create computational models for humour recognition or generation.

In this talk, I will address four important research questions related to
the recognition and use of verbally expressed humour, and I will bring
empirical evidence that computational approaches can be successfully
applied to these tasks. First, I will show that it is possible to
automatically construct a very large collection of humorous texts using a
novel technique for Web-based bootstrapping. Second, through experiments
performed on very large data sets, I will show how classification
algorithms can be applied to effectively distinguish between humorous and
non-humorous texts, with significant improvements observed over a-priori
known baselines. Third, I will illustrate how techniques for language
analysis can be used to uncover interesting properties of humorous text.
Finally, fourth, I will show how an automatic method for the selection and
addition of contextualized humorous text can improve the user-experience
and overall quality of widely used computer-based applications.

This is joint work with Carlo Strapparava and Stephen Pulman.