EducAIte

EducAIte front page and chatbot

Inspiration

Living in a diverse area of central New Jersey, many communities in our area do not have the funds or ability to access quality resources when it comes to tutoring and student support services. As all of us are the children of immigrants, the importance of education has been instilled in us from a young age and we believe that educational support should be accessible to everybody no matter their socioeconomic status.

What it does

The star of EducAIte is our machine learning chatbot that is powered both by our own training data, and by the OpenAI GPT-3 model. Once a user enters our site, they can proceed to ask for help in any of the supported subjects. These subjects are currently English, math, and history. Our chatbot will then use natural language processing to understand the user input and lead them to our machine learning model’s specific to the subject the user asked for. Our chatbot will then utilize both our training data, and the OpenAI GPT-3 model to answer the user’s questions, serving as a free online tutor. In addition, in our website’s menu bar we also have compiled high quality tips and resources to help students study more efficiently. We have also created a page that provides ways that students can access free or low cost tutors in their local areas.

How we built it

Our program is built using two machine learning models. The first one uses our own training data and a Tensorflow sequential neural network to determine the target subject for the OpenAI GPT-3 model. This model uses the entire internet as its training data, which gives it its conversational aspect and wide knowledge. Because of this, our personal tutor can provide quality responses to just about any question imaginable, all while having the conversational quality of a human. Our two models run on a FastAPI python web server, handling the requests and HTML templates. Speaking of which, our model is all displayed to our users with an inviting iMessage-esque design made from HTML and CSS.

Challenges we ran into

When building our model, we noticed some irregularities during tests. The model would randomly associate words with topics, leading to some strange results. We debugged for about two hours until we found the smallest mistake, a single indent. After correcting this indentation, our model worked very well, when creating our neural network, we noticed an indexing issue without code. It was a lengthy detour, but we eventually spotted our error and learned to always double-check our code.

Accomplishments that we're proud of

This program was a significant achievement for all three of us as developers. It tried us on several issues, which we persevered through, and in the end, we achieved a working tensor flow model, a link into gpt-3's ai model, a stunning UI that is easy to interact with and linking both the front and back end through the use of FastAPI, which now allows the program to work seamlessly. We are really proud of the fact that we blended so many different languages, libraries, technologies and resources to help create this software.

What we learned

We learned how to connect python to javascript through the FastAPI library, which now lets us connect a robust backend to a gorgeous frontend whenever we are developing other solutions in the future. We also learned how our code could serve to benefit people and be a source of fundamental change in the world. While it may only be a project now, the capabilities are endless with this code, and there is genuine good that can come out of this.

What's next for EducAIte

We believe that EducatAIte can truly make a difference in helping students across the country. We want to keep working on this software even after the hackathon to help add as many resources as possible. We want to expand the subjects that are chatbot is proficient in so that EducAIte can help the maximum amount of people. Since EducAIte is a one stop resources for helping students in their studies, we want to potentially contact schools and have them link our software on their website once we host everything publicly on the internet. This would help us attract students from all over and reach our target audience effectively. We would like to thank Ignition Hacks for prompting us to think of this idea, and we look forward to seeing what we can accomplish in the future through this idea.