Emojis Train AI to Recognize Sarcasm


"I'm being sarcastic." We've all had at least one exchange in which we either had to explain or had someone else explain that what was said was not intended to be taken straight. Generally, you need to know something about both the context and the speaker to grasp when to take a statement at face value or interpret it as sarcastic.

That's why it's particularly challenging to get handle on intent when attempting sentiment analytics on social media. For artificial intelligence to truly understand what humans mean, it needs emotional intelligence, as well. Iyad Rahwan, an associate professor the MIT Media lab and one of his students, who developed the algorithm with one of, Bjarke Felbo worked on just that.

The results are what they call Deep Moji. Described as "artificial emotional intelligence," Deep Moji was trained on millions of emojis "to understand emotions and sarcasm." Rahwan explained to MIT's Technology Review that in the context of online communication emojis take on the function of body language or tone in offering nonverbal cues for meaning.

The amount of data that went into the training was massive. They started with 55 billion tweets, which they narrowed down to 1.2 billion that featured one or more emojis from a list of 64 common ones.

The first part of the training was getting the system to predict "which emoji would be used with a particular message, depending on whether it was happy, sad, humorous, and so on," Technology Review reports. The sarcasm recognition was built on "an existing data set of labeled examples." The emoji training made the system more accurate at identifying sarcasm than algorithms that had not gone through the same.

The researchers put DeepMoji to the test, not just against algorithms but against "several benchmarks for sensing sentiment and emotion in text." They then tested it against humans, and it did exceptionally well. "It was 82 percent accurate at identifying sarcasm correctly, compared with an average score of 76 percent for the human volunteers."

It is rather surprising that it would outperform humans as one would consider the average person would still be more fluent in sarcasm than AI. It would be enlightening to learn how many people were involved and if their background or native language may have been a factor.

The site is meant to be interactive, and people who visit are encouraged to put in their sentences and label them. The video about it, that you can see below, ends with a call to action that people visit the site "to play with phrases and help turn words into emotion."

The additional input helps the system advance its learning and understanding of expressed sentiment. Visitors not only can enter tweet-like statement to see assigned emojis but can put in their own notes on the emotions behind them. Rahwan told Technology Review that self-identifying in that way is actually more accurate than having volunteers label other people's posts with the emotion they think is intended. Those fail to "'capture what psychologists would consider true sentiment,'" he insists.

Having come to read emojis, the system also generates them for the text put into it. I tested out some of the canned phrases already on the site and one that I typed it. I noted that the confidence level varies a great deal, from low to high. For the first two sentences I put in, the confidence level was high, as you can see here:

But I wanted to come up with something that shows some of the range and so then typed in one that only shows low confidence:

While seeing the emojis linked with statements may appear to just be a sort of modern day parlor trick, the purpose behind this emotional understanding is a serious one. The goal is to help combat hate speech. In fact, the researcher's original intent was to create something that would identify racist tweets. But the system needed to learn emotional context and sarcasm to accurately read tweets.

While the goal to improve the civility online is a noble one, improving machine-human communication is also helpful as an end in itself. With the increasingly popularity of IoT and voice-activated technology, we will have more and more people talking to their machines, and they will expect to be understood without extra explanations. To make that work, the emotional component of language has to be mastered by the machine.

Ariella Brown,

Ariella Brown is a social media consultant, editor, and freelance writer who frequently writes about the application of technology to business. She holds a PhD in English from the City University of New York. Her Twitter handle is @AriellaBrown.

Hiring Trend: Machine Learning Reduces Bias, Increases Applicant Pool

Unilever is using machine learning tools to increase the pool of entry level applicants and then test and evaluates candidates in the pool to cull those who are the best fits. Adding machine learning tools reduces bias, increases diversity, and makes the whole process more efficient, according to Unilever.

Analytics and School Attendance: A Laundry Story

Appliance maker Whirlpool suspected there was a correlation between student attendance and access to clean clothes. Here's the story of how the company placed washers and dryers in schools and tracked the difference it made in student performance.


Re: Sarcasm, irony, social
  • 9/20/2017 10:49:23 AM
NO RATINGS

@SaneIt So instead of using emojis as clues, it would use the laugh track and take in the differences between louder laughs and lighter chuckles. That seems possible.

Re: Sarcasm, irony, social
  • 9/20/2017 10:14:59 AM
NO RATINGS

Agreed that not all humor is going to elicit an audible laugh but as you mentioned at Applause and Laugh etc audience cues are the same concept.  Writers wanted the audience to engage at specific times.  These cues would work for an AI as well to teach it the subtleties of humor or any emotion really.  I was just thinking that we already have a lot of these cues built into television shows especially from the 80s and 90s so the reference materials are readily accessible, no need to add emojis to them.

Re: Sarcasm, irony, social
  • 9/19/2017 8:27:54 AM
NO RATINGS

@SaneIt it's an idea. However, not all humor is literally LOL funny. So it wouldn't necessarily inspire audible laughter. There are some verbal exchanges or pictures that you can appreciate as funny without actually laughing at them. BTW the sound effect of the laugh track is actually very old and predates television. I saw a movie from the early 40s in which audiences attended a radio program and were "encouraged" to laugh, applaud, etc. at the right moments.

Re: Sarcasm, irony, social
  • 9/19/2017 8:16:28 AM
NO RATINGS

The laugh track seems like the equivalent to using emojis in detecting sarcasm.  If an AI hear a laugh track after a statement that would be the cue to tuck some information away regarding the humor of the statement.  You could probably train an AI on 80s sitcoms and get pretty close to it detecting humor in the real world if you use the laugh track.  I'm assuming the AI would be able to go back and look at the context, inflection, intended audience and who delivered the joke.  At least it seems like humor would be easier to teach than sarcasm. 

Re: Sarcasm, irony, social
  • 9/18/2017 8:36:29 AM
NO RATINGS

@SaneIT an artificial laugh track for artificial intelligence, brilliant!

Re: Sarcasm, irony, social
  • 9/18/2017 8:35:50 AM
NO RATINGS

LOL @SaneIT you must be quite fluent in emojis. I almost never use them, except to soften something with a single one. I did make an exception, though, for tweeting this blog.

Re: Sarcasm, irony, social
  • 9/18/2017 8:25:29 AM
NO RATINGS

@Broadway, humor is difficult in any language even if it's your first language sometimes.  I think of watching Mel Brooks movies with friends and watching that one friend who is lost through every exchange.  It's hard to tell if they understood the joke and don't find it funny or if the humor just went over their head.  Teaching an AI humor is going to be tough and will probably require a laugh track to make it look like the AI is getting it.

Re: Sarcasm, irony, social
  • 9/18/2017 8:20:59 AM

@James Connolly, I think he would have written the whole thing in emojis and it would have become viral in today's culture. ☘️🚫🍞🚫🍖💸?  🍴👶

 

Re: Sarcasm, irony, social
  • 9/18/2017 8:05:47 AM
NO RATINGS

@Louis I was thinking of your comment when I saw I retweeted something today. We need an irony emoji for the irony in real news.

Saudi Arabia, distinguished member of UN Human Rights Council & UN Women's Rights Commission, welcomed by UN chief.

KSA Mission UN @ksamissionun
الوزير #عادل_الجبير يلتقي الأمين العام للأمم المتحدة بحضور سفير خادم الحرمين الشريفين لدى واشنطن والمندوب الدائم للمملكة لدى الأمم المتحدة.
12:21 AM - 18 Sep 2017

Re: Sarcasm, irony, social
  • 9/18/2017 8:02:33 AM
NO RATINGS

@Broadway there are many people who can't distinguish good from bad or cheesy humor. Much of that is subjective, so I'm not sure you can come up with any particular standard. Trying to program that into algorithms would likely reflect the biases of the programmers for or against certain types of humor.

Page 1 / 3   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +