Inspiration

My inspiration is maths Im a A level student who does the a level subjects maths further maths and computer science and honestly I've been into this topic neural networks for a couple months backpropogation and the linear algebra around it fascinated me

What it does

It allows neural network visualisation by the user clicking i would have prefered it be the user hovering but i dislike the amount of movement your mouse has to go through to hover

How we built it

by using css java script and html I planned on making it have delayed reactions which was my first project however i thought it was stupid id have rather make the nerual network logic the backpropogation and more that would have also took stupidly long so we created random values and made those values effect the size of the neurons fake neurons and created a fake loss to emulate backpropogation showing how loss near constantly decrease we looped a subtraction of 0.5 to a random value every time to show it decreasing

Challenges we ran into

it was to long the amount of div in my test run was way to challenging.I also had to look at my past algorithm because (sorry for my informality) during the holiday i had time to be cracked and built a neural network and i also pseudo coded a stock neural network this was on my github so i had to look back at it to see the logic and sort of figure out how to incorporate it

Accomplishments that we're proud of

What we learned

about machine learning but also how we visualise data

What's next for Neural netowrk visualiser

I hope to improve it but my skills as of now need improving

Built With

Share this project:

Updates