Inspiration

As students, we study using lecture slides, e-books and digital resources in the form of PDF files every day. This means opening your PDF files in one tab, then simultaneously searching keywords and topics that you don't understand in five other different tabs for a deeper understanding. Moreover, we are often studying notes and books written 20 years ago with no idea about latest developments and news on the concept. Digital Curriculum has taken over education resources and projections show the e-learning market worldwide is forecast to surpass 243 billion U.S. dollars by 2022.

This is why we want to enhance the e-learning experience with Learnwise.

What it does

Learnwise is a platform where you can upload your PDF resources, allowing you to read them in an enriched way while seeing definitions and news summaries related to the content in real-time at a glance. For example, when you open your machine-learning lecture slides, Learnwise fetches the keywords and key concepts in the PDF page by page, showing you the definitions and top related articles online with a summarized version of each one. You no longer have to go through the hassle of switching different tabs to search key concepts on every article.

How we built it

In the past 48 hours, we used React to build our web app. The web app takes in a pdf file and sends them for analysis to our flask server, which in turn talks to IBM Watson for our text analysis. Using Watson's Natural Language Understanding API, we get the relevant concepts and entities in each page of the PDF and use these concepts to get recent and relevant news articles using IBM Discovery News. Then we summarize these news article URLs using Agolo Summarization API to give a concise presentation of the news article.

Challenges we ran into

  • Designing the web app into a very reader-friendly interface.
  • Figuring out how IBM Watson works and running out of IBM Watson credits and creating multiple accounts to get more (life hacks).

Accomplishments that we're proud of

  • IT WORKS!
  • Solving a major problem for students. Learnwise can help you stay updated and have a more complete learning experience and we would love to use it ourselves.
  • Although we had basic knowledge in Natural Language Processing, we made a good use of the existing APIs (Watson and Agolo) to handle the NLU tasks instead of trying to re-invent the wheel. We believe that leveraging such technologies to handle great use-cases that can help the world is where our motivation lies!

What we learned

  • Watson is fast and awesome!
  • Our frontend developer learnt a lot of new things about React, specially how to handle files.
  • Our backend guys picked up Flask and new concepts about Natural Language Understanding.
  • This was the first hackathon for our designer and she learnt a lot about our stack and how web apps work!

What's next for Learnwise

  • Leverage Watson more, to filter out irrelevant content and add filtering by entities, concepts etc.
  • Make it faster! With multithreading and caching, the system can be much faster.
  • Make the PDF annotable and add ability to bookmark and save the definitions and news articles you like!

In the future, we want to extend the same concept and technology for video resources and all kinds of e-learning resources for students.

You can try the demo with the PDFs uploaded at https://github.com/hacknyu2019/frontend/tree/master/files

Built With

+ 10 more
Share this project:
×

Updates