Students studying in their mother tongue in the hinterland don’t get the same kind of access to digital resources that their peers in urban areas get to get their doubts cleared. In order to address challenge, computer engineers at the Indian Institute of Information Technology (IIIT-Hyderabad) have developed a voice-based technology solution that can provide quick answers to students’ queries.
They bring in a lot of artificial intelligence and machine learning that work in the background to help the students get answers in a few seconds.
“We are making it so convenient for students that they can use any phone to pose a question. The A multi-partner project spearheaded by the Raj Reddy Center for Technology and Society aims to supplement high school Science instruction in regional languages via an AI solution that operates sans internet.
For now, the team is working on the subjects physics and biology. “We will open it to other subjects later on,” Arjun Rajasekar, Senior Research Scientist, RCTS (Prof Raj Reddy Center for Technology and Society), told businessline.
When a student picks up a phone, dials the AskAgasthya and asks – ‘Kanti ante yemti?’ (What is light?), the solution converts the voice into text and translates it into English. It will interact with a relevant document in the data pile (extracted using the solution developed by Subtal.ai, a startup working in artificial intelligence technologies). After getting the answer, it will be pushed backwards using the same processes.
“The students will get the answers in a few seconds as the solution ensures a seamless process,” he said.
The team is currently developed the product for Telugu students and still is in the process of getting more voice samples to help the machine understand different ways of asking questions and in different accents.
The team is working with the Agastya Foundation, subtl.ai and a few other labs in the institute that have developed technologies such as NLP (Natural Language Processing) to build the solution.
Though the system has the ability to extract information from any given database, the team is restricting itself to NCERT and SCERT (state syllabus) curriculum. Also, they are targeting students of Classes 8 to 10.
Published on February 17, 2024
