COGnition

• 

  Home page of our website

• 

  Example generated summary

• 

  Example generated questions

• 

  Example generated flashcards 1

• 

  Example generated flashcards 2

Inspiration

We wanted to do something in the realm of AI and decided that creating a page where students could use LLM technology to create study tools like flashcards and summaries would be very useful, since creating tools like that for studying is an intensive task that only ends up giving the student less time to use the tools they made.

What it does

The project uses OCR technology to scrape text from an image or pdf. Then, it generates summaries of user inputted text such as notes or articles, generates questions designed to test the user's reading comprehension and understanding of the text's content, and generates flashcards of the important terms and definitions in the text in order for the user to study them.

How we built it

For the front-end, we used React.js, Bootstrap, JavaScript, HTML, and CSS. For the back-end we used Flask, Meta LLaMa, Python, Pytesseract, PIL, and PyPDF2. For the database, we used PostgreSQL.

Challenges we ran into

We had some difficulty in integrating the back-end with the front-end. We were all very unfamiliar with GET and POST requests, and synchronizing JavaScript and Python proved to be a challenge for us. However, we put our minds together and figured it out, leading to a impressive outcome.

Accomplishments that we're proud of

We are proud of making a product that was both interesting and practical. We were able to work together as a team and integrate all of our individual parts of the project. We worked with state of the art technologies such as LLMs, React,js, and PostgreSQL, which is something that we have never done before.

What we learned

We learned about the work that goes into making a website and integrating all of the various moving parts. We also were able to think creatively, thinking of new ideas and concepts along the way. Finally, we learned about our passion for computer science and building projects, and we strive to become better programmers in the future.

What's next for COGnition

We want to implement a cache-based login system, where users will be able to save their generated text and study tools based on their login credentials. We would also like to try using a more advanced paid LLM, since the free one we used for this project was a bit slower and more inaccurate than we would have liked. Finally, in the long term, we would like to develop a gamification system, where users can earn points for studying and potentially put them towards some kind of virtual reward, which we believe would increase productivity in students all around the world.

Built With

• bootstrap
• css
• flask
• github
• html
• javascript
• json
• llama
• llm
• pil
• postgresql
• pytesseract
• python
• react
• regex
• trello