Inspiration
We wanted to create an educational related platform as a result of the rise in online learning. Through our industry research, we found that children were the biggest victims of the COVID-19 pandemic in regards to their learning and that a lack of online learning tools available to them played a role.
What it does
Through machine learning (natural language processing), our web application known as Quiz.ai takes in educational text which users upload via PDF or images and then generates questions for user's to then answer. This allows user to immediately perform active recall and users are also given analytics of how they performed over time of using our service.
How we built it
The front end was made with JavaScript, and our application will take user uploaded images, PDFs, or pasted text, and generate questions into a quiz format for active recall. The generation of the questions is done through a combination of Google cloud vision API and Python natural language processing libraries. Google cloud vision API to detect the text in the image, followed by a summarization of the text to only keep the parts of the text. After we have decided which parts of the text to keep, we will convert all of the sentences into active voice, which can then be used to generate questions.
Challenges we ran into
Complication we ran into was trying to connect firebase to our pure JavaScript web app. Because we had not used NodeJs, it became complicated to be able to call our Firebase functions in our web app. In addition, Firebase functions have file size restrictions and because we have large training data files, we wouldn’t be able to run our back-end entirely through Firebase, which is why we have decided to write a web service using Python Flask and abandon Firebase functions, but to keep Google cloud vision API
Accomplishments that we're proud of
We're proud of the amount of learning that was accomplished, that we were still able to collaborate despite the challenges of not being in person, and that we were able to use our interests of machine learning in a positive way.
What we learned
We learned natural language processing, firebase, time management, and effective online collaboration skills. We also gained a better understanding of the challenges which children are facing during the pandemic.
What's next for Quiz.ai
- Our next steps for this product consist of developing more features such as a database to store generated questions which can then be accessed by other users, the flexibility for users to create their own tests from these different questions, students being able to formally test themselves in a structured way, leaderboards for different topics, setting up class tests as a tool for teachers, providing teachers analytics on how their students are performing
- Expand our team to gain a more diverse skillset in software development and education
- If we see our product being possibly commercialized, our team would look into joining a startup incubator as a means to validate our idea, develop and refine our idea, and acquire funding to ensure it’s success
Log in or sign up for Devpost to join the conversation.