Project Beats

Our home page
Our About us
Our Music Generator

Inspiration

Our team is inspired by the recent breakthroughs in AI technology and the capabilities of machine learning. Eager to get our feet wet in this field of CS, we decided to use machine learning in a simple yet fun way that is easily approachable.

What does Project Beats aim to accomplish?

The goal of Project Beats is to make machine learning approachable and give people a sense of what can be accomplished through AI. Anyone can do it.

What is Project Beats and how does it function?

The Project Beats website allows a user to create music using Machine Learning! Our user interface provides smooth interaction through three different web pages. The Home page features an up and down button that interacts as a wheel animation and allows the user to explore our hack's features. On the About Page, one can find information about the four hackers that worked together to make this awesome project. Finally, our Generate Music page is what it's all about. This page generates music through a machine learning program known as Magenta. Magenta is an open-source research project that focuses on music and art. Magenta's main purpose is to be used as a tool for machine learning ventures. In our case, with the little time that we had, we used their already trained networks to generate music on the fly and apply it through our website. Magenta offers both python and javascript, we headed towards more of the javascript path as it allowed for easy access of the midi files.

How we built it

Project Beats was built on the foundations of HTML, CSS, and JavaScript. The website is mostly HTML and is a major element of the project, and CSS was also used in front-end development. JavaScript was used to implement audio files into the website from Magenta.

Challenges we ran into

Our team ran into a few challenges over the course of this project, luckily, we have Santiago. Right of the bat, we had to decide what API is right for the scope of our project and what would be the most efficient machine learning network to utilize given the project's constraints. After deciding to use a website for our UI, we began to learn HTML and CSS for the first time. Over the course of this project, we have faced many challenges in front-end web development, for example, designing a functional and easily understandable web experience. We've also had to find ways around other challenges, such as implementing audio files into the website.

Accomplishments our team is proud of

Developed and designed a functional and aesthetically pleasing website
Created music generated through AI

Built With

Submitted to

Knight Hacks

Created by

Set up the website and had to endure the wonders of 404 and 403 errors. Edited and typed some code, README files, and the DevPost.

Jessa Ecle
Built interactive website using HTML/CSS/JavaScript,
effectively created user-friendly interface with custom features.

Christopher Nguyen
UX Engineer. No idea what I was doing initially but I figured it out along the way.

Daniel Rangel Garcia
I worked on the back-end with machine learning applications to have the computer generate the music through Magenta's Neural Network. Ran several different networks to get a good result, ended with a pre-trained network called improv_rnn. I gathered the midi that were generated and displayed them on the generate music page of the website.

Santiago Acevedo