AI learned to recognize sarcasm in Twitter on emoticons / SurprizingFacts

The creators of the "Simpsons", as in many other cases, predicted the appearance of this invention long before it was created by real scientists

Some human actions remain incomprehensible to the computer. But part of the behavioral features of computer systems have learned to recognize, and very good. For example, the IBM Watson cognitive system can determine the emotional tone of a letter. Teaching a computer to define emotions is difficult, but possible. The other day, another system was introduced, which is able to understand sarcasm. In this case, it's about sarcasm in messages from Twitter. The developers of this system claim that it can determine the emotional content of messages of different users better than in most cases it is done by the person himself.

Why is this all necessary? First of all, for companies to determine the attitude of users of social networks to their products and themselves. Now, key words and some other methods are used for this. But if the computer can determine the emotions of people sending messages, then this can significantly improve the efficiency of companies. In addition, if the machines confidently determine the emotions of people, it will help users understand what emotions used by another person who sent, for example, an e-mail message.

From the very beginning, the developers of the "emotional" algorithm wanted to create a system that could detect the posts of racist content on Twitter. But soon after the algorithm was ready, the project team realized that it gives many false positives. That is, the machine did not understand, for example, comic or sarcastic messages, and took everything at face value. So there was a need to teach AI to recognize at least sarcasm.

This algorithm uses deep training, which is becoming more common. As key indicators, experts decided not to use words, but smilies. Yes, they are now contained in most of the messages, so that you can roughly understand the mood of the tweet itself or another user's message on the social network. As soon as the researchers were able to achieve what they wanted, they realized that the AI ​​was working hard.

"Since in online mode we are unable to use non-verbal cues that help us understand what we are talking about, we started using emoticons. And the neural network was trained to understand the relationship between the emotion content of the tweet and the emoticons contained in this message, "said Iyad Rahwan, a processor from MIT Media lab, head of the research group.

The training of the neural network was really large-scale. In order for the algorithm to correctly determine the relationship between the emotional content of the message and emoji, the scientists collected a base of 55 billion tweets, then allocated 1.2 billion messages with emoticons (only 64 types of popular emoticons). Initially, they taught the system to predict which smiley should be used in a particular message, depending on whether it is cheerful, sad or some other. After that, the system was started to train to identify sarcasm by the presence in the message of already studied computer patterns.

As it turned out, the neural network learned to define sarcasm much better than the experts who developed it. The system correctly identifies sarcasm in 82 cases out of a hundred. The person, on the average – in 76 cases from hundred.

Neuronet was also trained to add emoticons to messages with a certain emotional mood. And with this task the computer also coped well. You can try the algorithm here on this site. Here you can help scientists to improve their algorithm by sending tweets equipped with the right emoticons. This makes the work of the neural network even more accurate.

Some experts who had time to get acquainted with the work of scientists, said that the use of smilies for learning a neural network and creating an "emotional algorithm" is a great idea. As for sarcasm, there can be a problem here – the fact is that not so many people are able to identify sarcasm. Some people do not even know what it is. Therefore, it is difficult to say how useful a neural network can be, which can determine sarcasm. But since it already exists, it means that somebody needs it. In addition, other projects more universal can be developed on the basis of this work.

About the author


Add Comment

Click here to post a comment

Your email address will not be published. Required fields are marked *