Now Computers Can Recognize Sarcasm on Twitter

You may find some totally genuine, deep-felt messages or comments full of sarcasm if you spend at least some hours in a day in social media such as Facebook, or Twitter. That makes us believe digital communication hasn’t diminished people’s sense of sarcasm yet. It’s no surprise if some researchers try to teach computers to identify when people are being sarcastic or not, and fortunately, they’ve had some solid success so far.

At a recent conference, two computer scientists, David Bamman from UC Berkley and Noah Smith from University of Washington presented a paper describing a system capable of recognizing premises of sarcasm on Twitter. They said their computers were trained to detect sarcasm based on a diversity of factors, including keywords (“shocked”, “gasp”) as well as exaggeration (“really”) and also hashtags like #lol and certainly #sarcasm. Most interesting part is while most of computational approaches to detecting sarcasm simply analyze the linguistics, the researchers here tried to include some wider context that made their detection methods much more reliable by correctly guessing 85 percent of the time if a post was sarcastic.

Detecting sarcasm may seem to be funny to you, but this could lead to better, more sophisticated efforts at tracking and filtering undesirable content from web comments or social media posts. A computer might be able to differentiate between when a status is meant to be joke and when it is meant to be serious. However, Bamman and Smith’s paper doesn’t get into the practical applications of their research. So, it’s a bit doubtful whether training computers to understand human sarcasm can go a long way towards making them smarter.