One thing digital communications doesn’t seem to have diminished is everyone’s sense of sarcasm. Spend a few minutes on Twitter, Snapchat, or in the comments section of this website, and you’ll find some totally genuine, heartfelt messages (yes, that was sarcasm).
Perhaps all these remarks could be put to good use, though. Some researchers are now trying to teach computers to recognize sarcasm in an effort to improve computers’ ability to make sense of human communications.
In a paper presented at recent conference, two computer scientists, David Bamman at UC Berkeley and Noah Smith at the University of Washington, describe creating a system capable of recognizing instances of sarcasm on Twitter. They developed their system using tweets that included the hashtag “#sarcasm” before testing it on new messages.
Previous efforts to automatically recognize sarcasm in text relied entirely on linguistic cues. What’s interesting here is that the researchers tried to include some wider context, such as who the author was and what they were tweeting about. And they found it to be noticeably better than existing approaches, correctly guessing 85 percent of the time if a post was sarcastic.
Being able to spot sarcasm reliably could prove very important in efforts to filter undesirable content from Web comments or social media posts. Humor is an important area for those trying to understand intelligence, too, precisely because it requires a pretty deep understanding of the way the world works—and how it doesn’t.
Indeed, a recent post by one of our authors describes an interesting effort to teach an algorithm to recognize humorous images, and even to create its own. That work is pretty crude and basic right now, but it points to an important new area of AI research.
So it might be a while before your computer can generate its own sarcastic jokes. I’d enjoy the lack of irony while you can.