Is Machine Learning magic? No, lets demystify it!

This is a "quick" crash course on machine learning
You will hopefully be able to use this video to start understanding and using machine learning today.

final project
This uses everything we talk about in the lesson part of the video. With this the goal is to allow you to walk away with something that can used and change to fit a bunch of different projects.
https://github.com/MatissesProjects/webcamKnn

fully connected machine learning visualization
Nice tool for seeing a fully connected neural network taking in your data live. With this tool it becomes somewhat clearer that an input gets flattened and broken down into different transformations. With the tanh being the one we talked about most (tanh link below)
https://www.cs.ryerson.ca/~aharley/vis/fc/

convolutional neural network machine learning visualization
Nice tool for seeing a representation of a convolutional neural network taking in your data live. Great tool for getting a feeling for how a convolution changes the space, and how it might be helpful to do pooling, a few concepts that we slightly touch on.
https://www.cs.ryerson.ca/~aharley/vis/conv/

visualization of the ml spaces - tanh
karpathy's visualizations are some of the most amazing for understanding how the space warps and transforms. This is a lot of what I am referring to when I talk about graphing on taffy.
https://cs.stanford.edu/people/karpathy/convnetjs/demo/classify2d.html

tensorflow playground: Great visualization tool, also allows you to try to solve a few different datasets using visual fully connected networks that you can see the graph weights and results for, great learning tool!
https://playground.tensorflow.org/

teachablemachine: tool by google which allows for fun interactions, final project is based on this. Main differences between this and final project is that it is single model. Also the final project allows for you to see how the whole thing works, so its hopefully clearer to see how to expand forward!
https://teachablemachine.withgoogle.com/

Gaugan: awesome paint like resource that we used to make cool scenes at the end
http://nvidia-research-mingyuliu.com/gaugan/


Extras!
Good view of convolutions, visual representation of different types
https://github.com/vdumoulin/conv_arithmetic

TONS of machine learning models!
https://tfhub.dev

If you have any questions feel free to find me on https://www.twitch.tv/matissetec where
I try to stream after work around daily, but realistically I do 1 or 2 long streams during the week. Or join us on discord for some games and for answers to a few of the stream games too https://discord.gg/ZUMxF6q3EZ