Archive | May, 2014

An Adaptive Learning Model for K12

20 May

My wife and I have been planning to launch a technology-centric non-profit organization to serve the underprivileged kids in poor schools in India. Much research and analysis needs to be done, and it’s going to take us a few years to get everything right. Fundraising, needed to purchase hardware, would take a lot of time too. However, the software, which would be an online product utilizing an adaptive learning model, can be launched sooner, so that we can iterate rapidly and keep improving it.

On a very basic level, for third-grade math, here are the requirements.

A database of math questions with these fields:

  • Question ID
  • Difficulty level
  • Question number
  • Options
  • Correct option
  • Feedback

Adaptive algorithm

  • A multiple-choice question with four options is served.
  • When an option is selected and submitted, the option is compared with the correct option.
  • If correct, the corresponding feedback is displayed.
  • if incorrect, the corresponding feedback is displayed.
  • When the Continue button is clicked, the system checks if the question was answered correctly.
  • If true, another question at the same difficulty level is served to make sure the previous answer was not a fluke.
  • A total of three questions at the same difficulty level is served.
  • If all are answered correctly, the next question would be at a higher difficulty level.
  • If any question is answered incorrectly, more questions at the same difficulty level will be served.
  • If two or more questions are answered incorrectly in succession, a brief tutorial will be displayed showing how to solve the question.
  • The tutorial will be followed by a question at the same difficulty level.

Though there are multiple interpretations of “Adaptive Learning“, I interpret it as a learning model that adapts or shows content based on the users’ current performance. It’s not a one-size-fits-all product, which are common these days though the technology has made rapid advancements. Great content and a great algorithm can be integrated with the right hardware to teach the kids effectively. It’ll take time but I think it’s doable.

A Technology-Centric Non-Profit Organization

14 May

Learning is hard. For most people. Because most of us are average learners.

Learning is hard also because the teachers in schools are not well trained, the classroom size is large, the socio-economic status of some students is low, there are not enough resources in the classrooms, and/or the students don’t expend enough time and effort outside the classroom.

In developing countries like India, there are thousands of schools that suffer from these problems. Surviving on minuscule government funds and poor management, these schools fail to teach their students effectively.

Classroom - India

Classroom in India

A technology-centric non-profit organization
There are many non-profit organizations in India and the US that serve the underprivileged kids. They are focused on training teachers and arranging for resources. Is it possible to build and implement a technology-based learning product for such schools when these schools lack even basic computers?

This is a question my wife and I have set out to find answer to. A few years ago, we’d decided to found our own education-focused non-profit organization to help poor schools teach their kids. This model requires two things:

Software/Learning product: The learning product can use an adaptive model to help the kids learn by creating a personalized learning path.

Hardware: I envision two options.

  1. Cheap 10″ tablets embedded in the desks to prevent mobility and potential damage. This model can be used in the classroom under the teacher’s supervision.
  2. Large touchscreens in kiosk-style stations. This model can be used outside the classroom in common areas to foster group learning and collaboration.

In the next post, I’ll elaborate on my adaptive learning model.

An Adaptive Learning Model for K12

Forecast, Don’t Predict

8 May

Predict: to declare or tell in advance; prophesy; foretell

In short, a prediction is looking in a crystal ball and is based on one’s knowledge and gut feeling. We all think we are very knowledgeable and can predict the future, and we all know how things turn out. Exactly the opposite.

Crystal Ball

Crystal Ball

A few examples

Apple: A prominent VC, Fred Wilson, recently predicted that Apple won’t figure among the top three companies by 2020 while Google and Facebook will. His logic is based on his knowledge of the hardware industry. We have to wait six years to see how true this prediction is.

The Beatles: “The Beatles have no future in show business.” — A Decca Records executive to the band’s manager, Brian Epstein in 1962

Harry Potter: “Children just aren’t interested in witches and wizards anymore.” — A publishing executive writing to J.K Rowling, 1996

iPhone: “There’s no chance that the iPhone is going to get any significant market share. No chance.” — Microsoft CEO Steve Ballmer, 2007

Computers: “I think there is a world market for maybe five computers.” — Thomas Watson, chairman of IBM, 1943

Personal computers: “There is no reason anyone would want a computer in their home.” — Ken Olson, president, chairman and founder of Digital Equipment Corp., 1977

Invention: “Everything that can be invented has been invented.” — Charles H. Duell, Commissioner, U.S. Office of Patents, 1899

Computer memory: “640K ought to be enough for anybody.” — Bill Gates, 1981

Big iPhone: “No one’s going to buy a big phone,” – Steve Jobs, 2010

Forecast: to
predict (a future condition or occurrence); calculate in advance

On the other hand, a forecast is based on cold, hard data. Removed from gut feeling or one’s own knowledge or emotion, a forecast takes lots of data and extrapolates from it. Could a forecast be wrong? Yes, it’s possible. However, a forecast always comes with a probability, the chance of something happening. While a prediction carries with it a heavy load of 100% probability, a forecast can be 0% or 100% or any figure in between.

Nate Silver, the founder of and the guru of statistical forecasting, has risen to be the most prominent figure in this field, especially after the 2012 presidential elections when many people and polls predicted a Mitt Romney victory. Nate Silver took the data from many polls, ran them through his forecasting analysis, and said that there was a 90% chance of Obama’s victory. His forecast of all 50 states came true. Not all of his forecasts for this year’s Academy awards were true, but at least, he didn’t make a fool of himself like so many famous people.

Suggested readings:
25 Famous Predictions That Were Proven To Be Horribly Wrong
15 famous predictions that were spectacularly wrong