N:zyme

Learn faster and more efficiently with N:zyme. Give your big blocks of text to our little enzyme friend, and he will transform it into a summary with pretty pictures for a more convenient way to study

Comment

Inspiration

Sometimes contents in a textbook can be dry and wordy. Is there a way to process this information like how enzymes process their chemical reaction in a more efficient and faster way? Introducing N:zyme, your little enzyme friend that will help you study the material in no time!

What it does

Import your textbook pdf, or copy and paste a chapter of it. Using language models and machine learning, N:zyme will compute the important takeaways for each textbook chapter in a concise yet detailed summary. The summary is also accompanied by relevant pictures to aid students who are visual learners. All the saved summaries can be used for N:zyme's active recall section. Active recall is a study method which demonstrated to be the most effective according to research on different studying techniques. Students will test their knowledge of the material by inputting keywords/phrases they remembered and N:zyme will compare it with its saved summary.

How we built it

Front-end: We started building a phone app using figma and mspaint (for the logo) to design the general layout of the website. We built the front-end of the website using React.js.

Back-end: Out backend was implemented in Python with the help of additional Python libraries, and API's. The Flask framework was used to implement the backend and connect it with the react front-end to retrieve client-side input data, while the Pandas python library was used to process and filter results from the NLP models our team utilized. The Co:here API's large NLP langauge generation models were used to generate text summaries and extract entities (keywords) from input text.

Challenges we ran into

Mild:

  • learning figma
  • learning react (HTML/JSX confusion)
  • github commits

SEVERE:

  • working with co:here api
  • working with flask (server error 500, etc.)

Accomplishments that we're proud of

We are proud to have a working prototype of this ambitious project within the dedicated time of 24 hours. We were also proud to be able to communicate as a team and learn from one another. Half of our team members are experienced in React which gave the opportunity of mentorship for the inexperienced members. Additionally, all of our members challenged themselves by using new technologies and being open asking for help, be it from mentors or from other teammates.

What we learned

As a team, we strengthened our React.js skills and learned how to use NLP models in our projects. Additionally, although some of our team members were familiar backend web development, none had used Flask before; as such, it was a fun challenge for some team members to learn to transfer their backend knowledge from Express.js to Flask in a short time. Outside of coding, our team learned the importance of efficiency and the importance of having a good idea for a hack (this allowed us to begin coding without wasting any time!).

What's next for N:zyme

What's next for N:zyme would be to implement voice recognition to text for learners who want to record their professors during lecture, as well to implement the ability for users to upload screenshots or PDFs of their textbook to automatically summarize the information contained within the media. We also want to work with the premium version of the Co:here API which allows more features and access to better NLP models for enhanced user experiences. Finally, we want to continue expanding on the "active recall" functionality of our application by adding an interactive section wherein users can quiz themselves on the summarized text by filling in the blanks, etc.

Share this project:

Updates