-
"Electricity" is linked to the relevant physics webpage
-
Inspiration
Studying for new subject is difficult and often confusing. Sometimes, we don't know even where to start. Our group thought it would be cool and helpful for students if computer recognizes problem types and lead them to right direction of solving them.
What it does
Our website takes in physics problem as text. Trained with machine learning, computer will be able to tell which chapter/section a question is from. Also, it gives quick link relevant page of Khan Academy, well known video lecture website. link
How I built it
We mainly used Python to perform machine learning. With easy and helpful library tensorflow, we could train the computer with pdf file of physics textbook. When user enters a question, the user input is connected to Python with flask and give answers back to user with good looking UI.
Challenges I ran into
As machine language require a lot of data set, our group fed textbook material to train it. But it was very time consuming and we did not end up with enough data to have sharp accuracy. It would have been much easier if there was a data set that we could use.
Accomplishments that I'm proud of
Although accuracy is not so great, our project is still functional. Everything from back end to front end and machine learning is all working. We are looking forward to getting more data after to enhance the accuracy of our program.
What I learned
Our group had no plans and we were suddenly struck with big project. We didn't end up getting new members in our team, but we still achieved a lot through HackDavis. Both of us feel more comfortable with newer technologies such as machine learning.
What's next for Chapter_Search
We are still looking for improvements in our projects. We want to serve students on other side of campus by supporting other subjects such as math or chemistry with our program. Furthermore, we are looking forward to learn image detection to use cell phone's camera and get input from user's working scratch paper directly to study guide.
Log in or sign up for Devpost to join the conversation.