eAIsy

Infographic for instructing user how to use software

Inspiration

We were inspired by a few of us having proficient machine learning knowledge. We want to make this knowledge more accessible, so we thought about how we could integrate machine learning into a web application so that the end user can input their data and not worry about the work.

What it does

Our project takes in user inputted datasets and uploads it so that it can be processed by a python script. The python takes the attributes of the data to dynamically create a neural net to run regression or classification on the data based on user input. The data is then returned to the front-end where it can be visualized in a multitude of ways.

How we built it

We created a web application using Next.js as the framework. DaisyUI and TailwindCSS allowed us to build the front end and with a reference to our python script, we were able to run the code necessary to generate and output the result of the model. The neural network was built using TensorFlow keras.

Challenges we ran into

Issues with using AWS services. Every service is unique and has its own limitations. Due to limitations in AWS Lambda, we would have to use server instances that cost more than we are able to pay for this page. Tying everything together was also an issue. Each part of the full stack infrastructure had its own nuances that made it difficult to have everything working in harmony.

Accomplishments that we're proud of

We're proud to have successfully created neural networks that perform with high accuracy. Allowing anyone to upload a dataset and learn underlying patterns by generating dynamically created neural networks based on user inputted code was a challenge that we were able to hurdle. Lastly, developing a front end that is aesthetically pleasing and user friendly.

What we learned

In working with new technologies, we each got exposure to other branches of computer science. We each had our own specialties and learning from one another was a educational experience. Those comfortable with frontend/backend/full stack got exposure to machine learning and vice versa.

What's next for eAIsy

Full AWS integration Convolutional Neural Network Support (Image upload) Natural Language Processing More generalized and sophisticated model generation