At our school, we had a huge problem with students telling the problems on tests in different periods. To try to solve this, teachers wrote separate tests for each period, but this forced many of them to take their personal time just to write different tests. We decided that this was unfair to the teachers and the students, and so we made an app to solve this problem.
What it does
Our app helps teachers by creating tests for them to print and give to their students. Then the teacher could simply take a picture of all the student's papers and it grades all of them automatically, using a deep learning algorithm to recognize the student's answers.
How we built it
We built this app using XCode. We tried to integrate Python into the Objective C, and we used Github to sync between our computers.
Challenges we ran into
Originally, we tried to make a webpage, but it didn't work at all, and we couldn't integrate a camera into html, so we decided to scrap it and make an app instead. When we started that, we ran into even more problems. At first, the Python integration was really hard and gave a lot of errors. Also, none of us knew any Objective C, so we had to learn that during the hackathon.
Accomplishments that we're proud of
We are really proud that we could actually create an app on iOS without any prior knowledge of app development. We were also able to use TensorFlow to develop an algorithm that could predict a number from an image with a 99.2% accuracy for the test cases. We were also able to create an algorithm that generates random systems of equations, linear equations and quadratic equations to solve. In addition, we had a set of AMC math competition problems to choose from. Our app can create a strong test and grade these tests quickly and with a high accuracy.
What we learned
We learned Objective C, python integration, and how to make an app.
What's next for Problem Session
Our app will be improved in integration of the TensorFlow and testCreater technologies with the iOS Objective-C. Also we will try to add different subjects, like Calculus, and maybe even English and History. We will definitely add a lot more problems to our data set.