Researchers at MIT (Massachusetts Institute of Technology) developed an AI system called DeepMoji that analyzes tweets on Twitter to detect Sarcastic comments, emotions, and sentiments and it works “better than most people.” The algorithm helps the computer understand sarcasm and emotional subtext so that it could fight racism, abuse, and harassment.
The AI system analyzes tweets on Twitter and detects sarcastic comments that most of the people could not even understand. Using such algorithm, computers can automatically spot and remove abusive statements and hate speech online.
Also, detecting the sentiment of social-media posts can help track customers’ attitude toward brands and products, and to identify signals that might point out trends in the financial markets. It can also be used in academics for research purposes such as to understand how information and influence flow through the network.
“What’s more, as machines become smarter, the ability to sense emotion could become an important feature of human-to-machine communication,” reads an MIT Technology Review report.
Initially, researchers were trying to develop a system capable of detecting racist posts but later they realized that without an understanding of sarcasm, the meaning of many messages could not be properly understood.
The algorithm uses a popular machine-learning technique called deep learning that relies on training a very large simulated neural network to recognize subtle patterns using a large amount of data. Researchers took advantage of emoji, a labeling system that many tweets use to represent emotional content and had a head start.
“Because we can’t use intonation in our voice or body language to contextualize what we are saying, emoji are the way we do it online. The neural network learned the connection between a certain kind of language and an emoji,” said Iyad Rahwan an associate professor the MIT Media lab who developed the algorithm with one of his students, Bjarke Felbo.
Researchers tested the algorithm against humans and found out that the AI system they have developed is better than the humans at spotting sarcasm and other emotions on Twitter. They also tested the algorithm dubbed as DeepMoji against several benchmarks for sensing sentiment and emotions in text and came to know that DeepMoji is far better than any existing algorithms at understanding texts.
“If machines are going to cooperate with us, then they’re going to have to understand us, and emotion is really hard,” added Rahwan.
You can also test the algorithm or interpret any sarcastic comment or slang at DeepMoji’s website, available for free to everyone.
Head over to developers’ blog post, paper, FAQs to learn more about the algorithm and how it works.