Computers are learning how to interpret your emotions

Some people try to convey emotions in their texts with punctuation, all-caps or emoticons. Soon they may not need to.

Eden Saig, a computer science student at the Technion-Israel Institute of Technology, has developed an artificial intelligence that can detect emotional sentiments expressed in a text message or email based on recognizing repeated word patterns.

That’s right. Your digital device will learn what you really mean when you send a text or email and convey it.

Saig has submitted a project as part of his course work, entitled “Sentiment Classification of Texts in Social Networks.” He not only got an A, he won the annual Amdocs Best Project Contest, sponsored by Amdocs, a provider of software and services to more than 250 communications, media, and entertainment service providers in more than 80 countries.

Saig said he developed the system at the Technion’s Learning and Reasoning Laboratory, a product of taking a course in artificial intelligence.

Lost nuances

When conversing with a friend, voice tone and inflections play an important role in conveying meaning. But in text and email messages, those nuances are lost and writers who want to show sarcasm, sympathy, or doubt have taken to using images, or emoticons, such as the smiley face to compensate.

“These icons are superficial cues at best,” said Saig. “They could never express the subtle or complex feelings that exist in real life verbal communication.”

Saig’s project collected all types of comments on web pages and catagorized them. Visitors were invited to submit suggestions for phrases that could be labeled as “stereotypical sayings” for that particular page.

When you look at these posts, Saig says certain patterns appeared. The method he developed enables the system to detect future patterns on any social network.

Teaching computers

Saig began collecting this colloquial, everyday language and soon realized the content could provide a good database for collecting homogeneous data that could, in turn, help teach a computerized learning system to recognize patronizing sounding semantics or slang words and phrases in text.

By using algorithms to analyze the content on web pages and using the results to automatically identify stereotypical behaviors found every day in social network communication, Saig was able to teach a computer to identify condescending dialog or slang.

The system focused on key words and grammatical habits that were characteristic of sentence structure implied by the content’s sentiments. Over time the computer learned to recognize the pattern.

“Now, the system can recognize patterns that are either condescending or caring sentiments and can even send a text message to the user if the system thinks the post may be arrogant,” said Saig.


You might be asking yourself what benefit could this provide. Saig says it has many potential pragmatic applications beyond clarifying emotions or feelings in interpersonal communication on social media.

At the extreme, he says it may help detect content that suggests suicidal ideas, what we may consider calls for for help.

At the very least, Saig says he hopes his invention can demonstrate to the writer how his or her words could be interpreted by readers, helping people to better express themselves and avoid being misunderstood.