Inspiration
When we were searching for a hackathon project, after pulling up the hardware list, we saw the muse. We were excited as we had never designed brain computer interfaces.
What it does
After calibration, this program uses a neural network to catagorize thgoughts into various terms (up, down, left, right, left mouse, right mouse) to control a computer through a "mental mouse"
How we built it
We built it using various elements from probability theory (such as the gaussian curve). We also created (from scratch) and used a neural network (in this project it was a feed forward 3 layer neural network).
Challenges we ran into
The largest issue we ran into was with muse itself. Many times it decided not to pair, other times it decided that port 5000 was in use (even though it was not). We use an the open sound control protocol wich was poorly documented on the muse.
Accomplishments that we're proud of
With that said, we were able program (from scratch) a neural net implementation with gradient decent. This allows us to perform machine learning with only our code.
What we learned
We learned all about Artificial intellegence, BCI, and the importanace of patience.
What's next for NeuralNet
We plan on using this NeuralNet we designed for our ImagineRIT project, an automated ASL interperter using the leap motion controller.
Log in or sign up for Devpost to join the conversation.