Skip to main content
search

Introduction:

India’s rural counterpart, Bharat, faces numerous challenges in basic numeracy and literacy. ML has tremendous potential to fix this malaise if used rightly.

Purpose:

“How we learn..” is a fascinating topic that the likes of Google are, researching all the time. There are numerous perspectives like environment, culture, teachers, etc. that one can look at. One of the perspectives is to keep the human at the center of the learning path, and that tools assist the human so that his/her grunt work is taken care of by automation.

Background:

“Practice makes the man/woman perfect!”. We think there is some truth to this axiom. The questions we asked are why there is no practice? Underfunding, both in Rupees and manpower, makes this a torturous process. Our well-funded school in Talasari taluka has 550 students from grades 1-10 with 6 full-time teaching classes. The astute reader will pick that there is not even a teacher per grade. Typically, classes 1-3 may be clubbed together, or an NGO funds a teacher partially, or the class practice loose to play outside. If you give a class of 50(conservative estimate) 10 sums to practice, what sane human will be able to check 500 sums the next day and give meaningful feedback? And that too when the teacher has pressure to complete the curriculum, act a as monitor for the next election or administer the next rubella or covid vaccine, or count the loos in the village for the Swach Bharat Abhiyan mission? These are tasks where automation can play a huge role. It lets the computational element do the generation and verification part, and lets the human intervention explain the concept where the student is stuck at.

Prior art:

Any solution has to meet  three vectors or has a large impact

  1. Is it scalable? (Can reach tens of thousands of students)
  2. Does it provide the learning gain that you wish to see?
  3. Is it cost-effective in Rupees and manpower terms?

We think ML has that potential when used appropriately.

Use cases:

Here are some use cases that enumerate ideas where ML can play a significant role –

  1. Improving multiplication tables via an audio recognition app
  2. Basic English alphabets recognition and writing skills
  3. Digital reading avatar for Read-a-story
  4. Automated chatbot on WhatsApp for practice

Prior art:

We have been experimenting in the tribal taluka of Talasari for the past 4 years. It has taught us:

  1. Human tutoring is critical to see a meaningful learning gain.
  2. The technological tool at hand has to be used as a tool, and not a teacher. Having a perfect tool does wonders.
  3. Development is not a linear process and you need Patience, patience, and, patience.
Author

Amod Joshi |  Director, Nplusone

Amod Joshi works as an educator to help with English and Math for the tribal children in Talasari taluka, Palghar. He co-founded Nplusone NGO which runs a Read-a-story program to help with English reading. In the past, he has worked in chip design and validation.

Close Menu