Guha, Kamlesh, Jayash, and Avaneesh appeared to be attending a typical computer science lesson at first appearance. The eleventh graders at Chennai’s DAV Boys Senior Secondary School are actually working to improve a piece of software they created earlier this year that aids autistic kids in understanding facial expressions.
All that was required was a laptop, some scratch code, and time spent teaching a computer to recognize the relationship between language and emotional states in people.
Due to repetitive thought and behavior patterns brought on by a neuro-developmental disease, children with autism have trouble communicating and interacting with others. These youngsters frequently struggle with seemingly easy tasks, such reading facial expressions.
Based on this idea, the team chose to train a machine-learning computer to recognize facial expressions in the hopes that it would then be able to assist children with autism in understanding non-verbal messages.
The concept is based on a straightforward yet essential premise. One of the students who worked on the project remarked, “First, we generated five labels to represent five emotions, and then we set about producing significant examples for these emotions.
“I really like the rain today, for instance, is a happy phrase. Hence, we filed it under “happy,” they continued. According to him, if someone types this line into the computer, it will categorize it as “happy” and assist the autistic person in expressing that emotion.
With time, the student group continued teaching the computer to recognize new statements as belonging to various emotions, such as “happy,” “anxious,” “sad,” and more, and to use “emojis” to express itself appropriately.
According to Jayash Suryawanshi, a fourth member of the four-person team, “Big tech companies supply these systems with billions of data when they develop machine-learning programs.” We could only feed it with 200 to 300 sentences because of time constraints, he continued.
The concept was developed by the group as part of a 21-day project for the Chennai-based EduTech portal Kruu. With 2.2 lakh students using its platform, the firm thinks there are countless opportunities for additional innovation.
Anil Srinivasan, co-founder of Kruu Inc., said, “I couldn’t be more happy that these are students who are 14 or 15 years old, who without being informed what empathy is and without being told what it means to labor for the good of other children, have done it spontaneously. That was made possible by technology, he continued.
The boys are not finished when they return to school. They are currently developing a Python version of their machine learning tool, which will enable them to handle and input much more data much faster and with more accuracy.
It’s interesting to note that the project’s success hasn’t necessarily meant that the four are actively considering starting a tech company. For Guha, Kamlesh, Jayash, and Avaneesh, the goal is still to use technology to connect with underserved kids.
