Artificial Engineers

Inspiration

What it does

The app can generate quizzes from a database of AI annotated CT scans to test students on whether they can identify what kind of condition the patient is suffering from. There is a user sign-up so that progress/overall score can be tracked for leaderboards. To make learning more fun, there is also a duel mode, where two users can play against each other to see who knows more.

We envisioned our app to be used both for fun, as well as in a classroom setup, in order to possibly integrate into learning platforms (such as moodle)

How we built it

For the backend, we used flask to build a REST API. For the frontend, we used react. To train our machine learning model, we used the Azure ML studio. For storing user data, images, login, etc. we used dynamodb and s3 bucket from AWS.

Challenges we ran into

The training of the machine learning model turned out to be somewhat challenging, as there are many classes associated with the dataset and the encoding of the labels was inconsistent between the test and train set. Furthermore, even on the reduced size dataset, training took a significant amount of time. We also had some challenges in getting the app deployed.

Accomplishments that we're proud of

Built a working MVP with all the features, we set out to implement in it. Decent login/user tracking with constant authentication checking, bearer token. Duel mode.