The Fifth Elephant 2019

Gathering of 1000+ practitioners from the data ecosystem

Tickets

Building a personalized learning system using a concept graph, and latest research in cognitive science

Submitted by Nilesh Trivedi (@nileshtr) on Monday, 25 February 2019

Session type: Tutorial

Abstract

I have been actively learning new things (beyond what was required for my formal education) since I was a teenager. A few things I have learned in this time are: mathematics, engineering, economics, philosophy, public speaking, a dozen or so musical instruments, a dozen or so programming languages). But the list of things I am yet to learn is not getting any shorter. I realized that I had to get better at learning itself, and started reading about research from cognitive science. I started exploring mathematics with my toddler which forced me to think deeply about how we learn and how we can get better at it.

After almost a decade of doing this, I started building a personalized system for learning anything in the optimal manner. This involves ideas like spaced repetition systems, chunking, meta-cognition. This system heavily relies on data (in particular, links to good pedagogical content) and the relationships between them (i.e. taxonomy & concept graph). I incorporated ideas from psychology (gamification, learning styles) and data science (summarization, recommendation) so that this knowledge discovery platform allows you to understand any topic through most efficient paths.

In this talk, I will explain the architecture of this personalized learning platform.

Outline

  • My journey
  • History of learning systems
  • Learning how to learn: Research in cognitive science:
  • Spaced repetition systems
  • Chunking
  • Growth Mindset
  • Focused vs Diffused modes
  • No speed limit
  • Learning styles
  • Building a concept graph
  • Curating pedagogical resources
  • Personalization
  • What you already know?
  • Multimedia
  • How much time do you have?
  • Data science:
  • Summarization
  • Recommendation

Speaker bio

I have been learning things deeply and making things for almost 15 years now. I completed dozens of hard MOOCs in mathematics, computer science, economics, psychology etc. I have used this knowledge to build useful things - be it music, apps, electronic/robotic appliances (like automatic portable roti making machine) or DIY-furniture. I have also been exploring mathematics with my 5-year old child, which forced me to think deeply about how people learn and how we can get better at it.

Links

Slides

https://docs.google.com/presentation/d/1n-tkTbF1k96wAiAsIl1Tk7B6Co83lSHMV-X9tjry_Nw/edit?usp=sharing

Comments

  • Anwesha Das (@anweshasrkr) Reviewer 9 months ago

    Thank you for submitting this proposal. We require preview video by 11th March, latest, to evaluate your proposal and make a decision.

    • Nilesh Trivedi (@nileshtr) Proposer 9 months ago

      Hi @anwesha,

      Please see the link given under “Slides” section. For some reason, the deck embed is not working properly.

  • Anwesha Sarkar (@anweshaalt) Reviewer 8 months ago

    We are looking into the problem, will get back to you. Meanwhile updload the preview video by 21st March(latest).

  • Zainab Bawa (@zainabbawa) Reviewer 8 months ago

    Nilesh, thanks for the submission. Who is the target audience for this talk? Who should listen to your journey and learnings? Where are these learnings applicable, beyond building a personalized learning system.

    • Nilesh Trivedi (@nileshtr) Proposer 7 months ago

      The audience for this would be datascience beginners who’d like to learn how to build a real-world application using techniques such as knowledge graphs, concept graphs, entity-relation modeling, inference engine etc. The domain modeling part was especially tricky in this case, and I believe my learnings are worth sharing.

      My project attempts to realize Danny Hillis’ idea of a Learning Map as he introduced here back in 2012: https://www.youtube.com/watch?v=wKcZ8ozCah0

      • Zainab Bawa (@zainabbawa) Reviewer 7 months ago

        Noted. Thanks.

Login with Twitter or Google to leave a comment