Misnomers

Inspiration

Both of us are big fans of all things fantasy, whether that be novels, video games, or tabletop RPGs. We thought it would be fun to see what an AI could come up with for unique names.

What it does

Misnomers generates unique names using a recurrent neural network (RNN).

How we built it

Misnomers is built 99% out of Python. Tensorflow was used to train the RNN, and flask was used to host the web app and API. Classic HTML/CSS was used on the front-end for a minimalist interface, with a small bit of JS to handle API calls.

Challenges we ran into

This project was a baptism by fire. Pretty much this whole project has been new experience after new experience. Before the past day and a half, neither of us had ever touched web development, or AI/ML.

Accomplishments that we're proud of

We are extremely proud that we even have a website that works in some capacity. Since we were both green at almost every component involved here, it's a feat that this site is even functional.

What we learned

We broadened our development horizons by miles. This project has been an excellent introduction to minimalist web development, web hosting, and machine learning. It has ignited a fire in us to try more projects with these technologies in the future.