Month: October 2013

O(log n) algorithm for education

The amount of time required to train a certain number of students in some skill is usually proportional to the number of students.

In this post, I wanted to talk about a process for training people using ‘social networks’ that can work much faster and better.

The method was developed by a social start-up called Digital Green from Bangalore.

Digital Green’s mission is to educate farmers and help them improve their farm yields.

There are about 263 million people in India who depend on farming and many of them are illiterate and uneducated.

Therefore, education for farmers, if done right, can make a real difference.

Government agricultural research institutes have, for a number of years, been evolving farming best-practices, and trying to educate farmers on the same.

The government’s method of choice for educational initiatives is rural field officers who do the rounds of the villages.

But the process is often ineffective in driving change on the ground.

Digital Green discovered the reason for that.

It turned out that farmers were hesitant to abandon or change practices that had worked for their ancestors for thousands of years.

They were finding it difficult to trade time-tested ways of doing things for new methods.

Digital Green discovered that what farmers found more convincing was recommendations from early adopters from nearby villages.

They were less resistant to ideas from persons they knew.

So, Digital Green went about using their insights to change how education was delivered to farmers.

They started by constructing a social network without computers.

They made video CDs of early adopters engaging in a farming practice that needed to be promoted.

They then mailed the CDs to nearby villages – in each of which television and video equipment had been installed – and showed the videos to people there.

The farmers watching the videos would see familiar faces, places, soils and crops, and might also be tempted to adopt the practice in order to be able to star in the next video.

So, the practice would spread virally, from one early adopter to his connections, and from them to their connections, and so on.

This model of training is very similar to the model of information dissemination in social networks. Like in many other forms of viral transmission, the amount of time it takes to reach out to a certain number of people grows logarithmically with the number of people. In other words, it scales very well.

So, Digital Green’s approach is an O(log n) algorithm for education.

The following BBC article on Digital Green contains an interview with its CEO Rikin Gandhi:  http://www.bbc.co.uk/news/technology-23867132

You might also like this article we wrote in 2012 on another O(log n) algorithm for educating large numbers of people:  Is there an algorithm to combat poverty?