I planned on writing more about the learning videos and material. However, there was much more content and work to cover prior to the first project due date than I anticipated. Some of the material covered includes: multilayer neural networks, activation functions, gradient descent, and back propagation.
In this project, I implemented a neural network with a single hidden layer (i.e. 1 input layer, 1 hidden layer, and 1 output layer). I trained the network and tuned parameters like the number of hidden units, number of epochs, and learning rate to produce reasonable training and validation errors. Overall the system did a decent job at predicting bike share usage.
I’ve been very interested in alternative education options like Udacity, Treehouse, Code Fellows, Coursera, etc. I speak about them often to coworkers and friends. They probably think I’m working for them. Currently, I’m not employed by any of those companies, but I’ve definitely looked at job postings more than once. 🙂
I have a background in computer vision/machine learning/robotics from traditional universities. I’m very intrigued by the low cost options, modern content, and relevant skills that many of these companies offer.
I’ve done a lot of work on some of the websites of the aforementioned companies. However, yesterday I enrolled in my first program at Udacity. I’m excited to learn about deep learning and experience participating in a Udacity program. While participating I’ll try to post relevant topics from the program. Hopefully these posts will remind me of what I’ve learned and can be used for my reference.