My Journey into Machine Learning and Neural Networks
The past few months I kind of went on a tangent trying to learn all I can about Machine Learning and Neural Networks. I had felt that I was being left behind on this new emerging technology. So I decided to play catch up.
I found a lot of great material along the way and thought I'd compile a list here of some of the things that especially helped me (Because I remembered them in particular). Hopefully this can kind of get you started in your journey to understanding Machine Learning.
Youtube
MIT Artificial Intelligence
Actual Lectures from the course, general topics of artificial intelligence and machine learning
Stanford Convolutional Neural Networks
Actual Lectures from the course, More on neural nets and present day tech
Siraj Raval How to Simulate a Self Driving Car
After watching this video, neural networks and how to use Keras really started to click for me.
This is the Source from the Video
Executable Downloads for training:
Books
I picked up all for volumes of the Handbook of Artificial Intelligence
Artificial Intelligence - Patrick Winston (Lecturer from the MIT Lectures on Youtube)
Learn by Doing
Kaggle is a pretty cool website that allows you to try to solve different machine learning challenges.
A kernel that I made up why trying to understand how they worked
Two pretty good challenges to start off on
This is a good way to kind of start understanding how to interpret and process data, and configure it in a way to be useful in machine learning.
A couple kernels that were helpful in getting me started to understand processing data
Good way to start understanding Image Classification MNIST dataset
Udacity Online Courses
Computer Vision
Learn more about image processing and detect things with in an image
This helped get my head around what is possible in OpenCV
Artificial Intelligence for Robotics
This is the first course I started to take on Udacity. It was a little over my head at the time, but now I think I'd get it. I want to restart this one
Things to Install
Python, you are going to need to know most likely, though there are definitely different options. I personally like python
Keras is a layer that makes using Tensor Flow make more sense. You can build networks without the nitty gritty
Tensorflow is the backend software that trains the neural networks. There are a few other options out there also, but I just picked this one first, so that is what I'm going to stick with.
OpenCV is a great Computer Vision, image processing library that is available in python. You can process images and live video feed from a webcam
Numpy basically helps manipulate arrays, and is needed for OpenCV
Pandas makes reading in CSV values and building tables easier
Comments