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
  • 9/30/2017 7:07:30 PM
NO RATINGS

I think you're onto something! These are great suggestions. I'll get right on it...

 

[partial sarcasm at no extra cost]

Re: Sarcasm, irony, social
  • 9/29/2017 10:00:25 AM
NO RATINGS

@SaneIT  I suppose I haven't watched that kind of children's programming to make the connection. But you know that -- like Harold Bloom argued for poetry -- films influenced each other and TV. The famous chocolate factory skit in I Love Lucy also had a precedent in Charlie Chaplin's Modern  Times, though he lacked the opportunity to stuff the candies into his mouth in his situation.

Re: Sarcasm, irony, social
  • 9/29/2017 6:58:22 AM
NO RATINGS

@Ariella, I suppose part of it is the fact that it worked before.  I think the multitude of movies like Batman, Spiderman, etc. are less direct re-creations of another work.  What I'm referring to is not re-using characters or a theme but nearly a shot for shot remake of a particular scene from an older television show.  I see it quite often even down to a tight shot of the actor's face as they attempt to mimic one of Lucille Ball's dramatic faces.  If you have kids you may have seen these and not even realized why they felt familiar.  I can't find good quality video of the copy cats but the candy factory scene from I Love Lucy's Job Switching episode has been done by several kids shows that I know of.  Two that are easy find on YouTube but don't have great footage are a show called Drake and Josh that my kids watched often as well as the animated version of My Little Pony.  I think it would be an interesting skit to train an AI with.  If it was able to watch the scene and pick out the funny bits in each version and recognize where it had seen them before I think that would be a good step toward a humor bot.

Re: Sarcasm, irony, social
  • 9/28/2017 4:07:45 PM
NO RATINGS

<I can't recall seeing an attempt at a remake of classic comedy scenes, so I can't comment on the quality of that. I will say that the original episodes of shows like I Love Lucy, The Honeymoooners, various Marx Brothers movies, Burns & Allen routines, etc. had a finesse and subtely of staging and comedy performance by highly skilled comedy actors that made the comedy work and become comic masterpieces.>

@Lyndon_Henry and people do hope to capitalize on that. See .broadway.com/buzz/189430/laura-bell-bundy-leslie-kritzer-michael-mcgrath-michael-mastro-to-lead-world-premiere-musical-the-honeymooners/

Based on the iconic CBS television series, The Honeymooners features a book by Dusty Kay and Bill Nuss, music by Stephen Weiner and lyrics by Peter Mills. This world premiere musical features direction by Tony winner John Rando and choreography by Emmy winner Joshua Bergasse with musical direction and vocal arrangements by Remy Kurs.

I'm just not sure that the show would ever have put in this plot:

In The Honeymooners, Ralph Kramden (McGrath) and his buddy Ed Norton (Mastro) are back and still shooting for the moon. After shocking their wives (Kritzer and Bundy) by winning a high-profile jingle contest, they are catapulted out of Brooklyn and into the cutthroat world of Madison Avenue advertising, where they discover that their quest for the American Dream might cost them their friendship. 

Re: Sarcasm, irony, social
  • 9/28/2017 4:04:51 PM
NO RATINGS

@SaneIT <  I assume it's because younger kids will not have seen these sketches before so it feels fresh to them.  > Also they assume if it worked before, it should work now and why bother to create something new when you can just play it safe that way. It's that same kind of thinking that gives us multiple film versions of Batman, Spiderman, and various Avengers.

Re: Sarcasm, irony, social
  • 9/26/2017 8:58:09 AM
NO RATINGS

I'm going to say that not much of it will translate well to AI given the fact that watching the young actors reproducing these scenes.  If humans can't watch them and re-create them without it looking painfully awkward I'm not sure that AI will understand the subtelties either.  If you don't have young kids around you'll probably miss these since it seems to be the shows for pre-teens that have latched onto the older shows for their material.  I assume it's because younger kids will not have seen these sketches before so it feels fresh to them.  

Re: Sarcasm, irony, social
  • 9/25/2017 4:47:41 PM
NO RATINGS

..

SaneIT writes

... I believe that what you're noticing is the formula for television in general.  I realized watching shows with my kids that many of the situations, story lines, jokes, etc. all came from shows I could marginally remember seeing as a kid.  I lost track of how often a kid's show did a skit from I Love Lucy.  They were almost a shot for shot remake in many cases.  I think we're seeing the great recycling of comedy right now. 

I can't recall seeing an attempt at a remake of classic comedy scenes, so I can't comment on the quality of that. I will say that the original episodes of shows like I Love Lucy, The Honeymoooners, various Marx Brothers movies, Burns & Allen routines, etc. had a finesse and subtely of staging and comedy performance by highly skilled comedy actors that made the comedy work and become comic masterpieces.

Now, how much of that can be effectively translated into AI/ML algorithms?

..

Re: Sarcasm, irony, social
  • 9/25/2017 8:32:57 AM
NO RATINGS

@Lyndon_Henry, I believe that what you're noticing is the formula for television in general.  I realized watching shows with my kids that many of the situations, story lines, jokes, etc. all came from shows I could marginally remember seeing as a kid.  I lost track of how often a kid's show did a skit from I Love Lucy.  They were almost a shot for shot remake in many cases.  I think we're seeing the great recycling of comedy right now.   Writers grab a simple concept that they know appealed to a wide audience and they adapt it to their characters then write an episode around how to get them into that situation.   TV has become a passive activity in many homes.  In my the TV tends to stay off most of the day unless someone sits down to watch some thing intentionally but when I visit friends or family the TV is always on even if no one is watching.  It has become background noise so the over the top and in your face programming feels like an attempt to capture eyes and attention. 

Re: Sarcasm, irony, social
  • 9/24/2017 9:24:24 PM
NO RATINGS

@Lyndon_Henry that's an interesting proposition, trying to direct the direction humor rather than serving up what is currently popular. The problem is that generally producers seek to go for the obvious choices in deciding on what would sell. That accounts for the constant remakes and actual comic books serving as the basis of a major franchilse of films -- all designed with multiple sequels in mind -- today. 

Re: Sarcasm, irony, social
  • 9/24/2017 6:11:51 PM
NO RATINGS

..

Ariella writes that

... humor can get very nuanced and subtle. I know my daughter claims that she doesn't care for the comedy in older films because it's filled with what she calls "connect-the-dots" type of humor rather than the more obvious jokes.

Several decades back I began noticing what I perceived as a decline in the quality of humor, especially as presented in the TV medium such as sitcoms. Humor was becoming far less incisive and reverting to a kind of comic book or animated cartoon level. It came across as "childish" ... but for adult audiences.

More nuanced and subtle humor fortunately still exists, at least to some extent in shows like SNL, some of the latenight comedians, and mock "news" or commentary shows such as those of Trevor Noah and John Oliver.

Hopefully a savvier new generation will bolster the sudience that appreciates this type of humor. AI? Remains to be seen ...

..

Page 1 / 5   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +