Soon, Google AI tools to help customise learning for the students

Bengaluru-based education technology startup Cuemath is months away from implementing Artificial Intelligence (AI) and Machine Learning (ML) tools to help it customise its curriculum for every single student on its platform based on their strengths and weaknesses.

The Google-backed firm, which focuses on teaching mathematics concepts to primary and secondary school students, will use some of the Mountain View firm’s open source ML tools to customise learning on its platform.

“We have already taken a few steps in this direction, building some adaptive capabilities into our product. Over time, more of our product will individualise the curriculum based on a learner’s specific needs and this has been at the heart of our philosophy,” said Manan Khurma, founder and CEO at Cuemath.

Individualisation of curriculum will ensure that all students are able to grasp what is taught to them, something which is essential in large-scale digital teaching. 

While Cuemath uses a hybrid model of ‘teachers plus technology’, it says even in a classroom of 15 students a teacher alone will not be able to customise learning for all.

The AI tools at first could help alert teachers of a particular student’s weakness, allowing them to dedicate more time or suggesting further problem solving or reading. 

In time, the AI will begin altering the curriculum shown to each student on its own, making for a truly customised learning experience.

“As of now, we’re extending our current technology stack to do this. A lot of data pipelines don’t exist and we’re building those to pull the data from the student into our system so that we can analyse it,” added Khurma.

Cuemath last year introduced comics that focus on teaching students math through stories. While the company picked general topics of interest based on demographic studies of its users, it says it would like to customise learning even further.

For example, students who might be more inclined to the sport of cricket, could be offered questions pertaining to math using cricketing statistics. The same concept could be taught to another student using a completely different topic. 

The trick will be to train the AI to not just figure out a student’s likes and dislikes, but also customise the curriculum based on what it finds.

“It’s clear that at scale we can’t drive this process by manual interventions, it only makes sense that this individualisation process is system driven. So at the heart of our learning system, we’re building an adaptive engine using all the latest technology tools. This is to ensure that every child’s curriculum is customised or personalised in some way,” says Khurma.

Cuemath is looking at growing its base of users to 50,000 this financial year while growing the number of centres it has from 2,500 to 4,000 by the end of the year. 

The company local tutors to partner with and says it has been seeing traction from stay-at-home mothers who have had some kind of background in academia. Apart from its focus on building a strong curriculum to teach mathematics to students, it also invests a significant amount of resources in training teachers. 

This is unlike many other edu-tech firms who are trying to do away with teachers altogether, owing to the added cost as well as inability to find a large number of trained teachers.For Khurma, while he admits finding good teachers is hard, he says they continue to serve an important role in a child’s education. 

Moreover, while the perception of parents that their children can learn using technology is slowly changing, having a more conventional setup of a teacher and student does help convince more parents to sign their kids up to Cuemath’s service.