AI For Good: The Shortest Path To Mastery Is Paved With AI

India has 315 million students – that’s a little short of the entire population of the United States of America! Generally, schools, government or private, have an upward of 40 students per class. In most rural areas, schools usually have just one teacher managing an entire school.

India most definitely faces multiple challenges in the education sector – the general upkeep of schools is a Herculean task, in addition to ensuring students receive quality education. Educators are at a loss for they cannot precisely identify the learner’s capabilities, due to gaps that exist across domains and grades.

More than 50% of learners in Grade 5 cannot read a Grade 2 text or solve a simple subtraction problem. Marks are not an accurate portrayal of what a learner knows and doesn’t. Moreover, a one-size-fits-all approach does not work with learners, who are at different levels due to learning gaps across domains and grades.

Former Amazon employee Cibe Hariharan and Teach for India alumnus Sethuraman TA, whose paths crossed in a classroom, realized just how fragmented the learning process is in India.

Sethuraman, who was a teacher in a government school, was having trouble gauging how well his students really understood concepts. Sethuraman recalls a student he had named Hari, The eight-year-old’s performance in standard classroom tests were less than average but Sethuraman could not detect what his problem area was. After spending half an hour with Hari, asking him a range of questions on his skills and challenges, Sethuraman found out that Hari didn’t understand three-digit subtraction and related concepts. It was likely that other students faced obstacles in other areas of mathematics, but was it practical for Sethuraman to spend 30 mins on each student to zero in on their problem areas?

Definitely not. And Sethuraman is just one teacher in India.

Hariharan and Sethuraman decided to leverage the power of predictive analytics and AI to chart a student’s journey at a microscopic level. They founded Jungroo Learning in 2018, a B2B edtech SaaS platform for educators. Jungroo locates the learner precisely, and creates the shortest learning path to mastery for them. It essentially mimics a teacher to generate micro-level mastery information of students. Students generally take up assessments suited to their personal competency levels. Based on this information, a customized path is created for every student. Moreover, teachers are provided with customized teaching plans for every student and group for future lessons. Jungroo even provides reports based on the needs of different stakeholders like parents, teachers and organizational heads.

Technology Powering Jungroo’s AI platform

A look at Jungroo’s technology approach reveals that it predicts the knowledge state in the learner, before developing a path for easy understanding of subjects. This approach is based on the Markov Decision Process, which predicts knowledge state with minimal student interactions. In order to predict the shortest learning path to mastery, Jungroo’s creators have modelled this concept as the Discounted Markov Decision Process with modifications in actions and rewards.

In addition to basing its intelligence on a student’s knowledge state, Jungroo Learning is constantly picking up new information from a student’s ongoing engagement on the platform as well. Jungroo uses the time spent, percentage watched, type of response, the correctness of the questions for a student and the relative performances, interactions of the class amongst other parameters to calculate the engagement score of a student. This is a very crucial parameter to do effective personalisation and adaptivity, keeping in focus the engagement and mastery of the student.

Jungroo Learning has developed a suite of products, which include an adaptive assessment engine, an adaptive learning engine, Automatic Question Generator and a low-cost real-time classroom polling system.

Adaptive Assessment:
Essentially, Jungroo mimics an oral examiner using an adaptive assessment engine. It generates insights based on information about the student at a very atomic level, and helps an educator pinpoint the exact problem areas for a student.

Adaptive Learning Engine:
The platform’s adaptive learning engine creates personalized learning paths using insights from the assessment engine. At the platform’s backend, Jungroo calculates the engagement score based on several parameters, and arrives at the shortest path needed to understand the concept at its core.

Automatic Question Generator:
Jungroo has an a​utomatic question generator​which can automatically create any number of questions spontaneously, based on predefined templates.

Classroom Polling System:
There is a real-time classroom polling system, which can be used in spaces where device and Internet access is limited.

Awards and Accolades

Jungroo Learning was inducted into the first cohort of InnovatED – Teach for India’s national incubator for entrepreneurs seeking to build impactful organizations in the education sector. They are also among the few selected for the Education Innovation Mentorship Program of Read Alliance, supported by USAID.

Jungroo Learning has also been chosen for a grant by ELEVATE – another startup-oriented initiative for the Government of Karnataka.

Jungroo Learning was one of the 10 finalists of AI for Good, conducted by NASSCOM CoE DSAI in association with the Government of Karnataka.




Share This Post

Leave a Reply