Hi all! Normally I try to update my blog at least 1-2 times a week with anything I find interesting. It has been harder this last week because I’ve been knee deep in the Facebook Udacity AI Scholarship Challenge 2018. 10,000 people from around the globe got accepted to this scholarship and 300 will move on to Phase 2. How do they choose who moves on? Well it is based on homework completion, a lab challenge project and community participation but they don’t tell you exactly how things are weighted. I’ll still try to make quick updates, especially when I need a break from AI.
When does it end?
It is a 2 month long course. Hopefully after Jan 9th 2019 I’ll be 1 of the lucky 300 to move on to the next course.
How hard is it?
It is pretty challenging if you don’t have any experience in coding. There’s a lot of linear algebra and calculus level math. But once you get past that, you’ll realize that the PyTorch framework (created by the Facebook AI Research Team) makes it very easy to define neural networks with all of their built in libraries. Most of the math functions are already made for you, it’s just a matter of manipulating and pre-processing data so you can toss it into the neural network.
Would I recommend this course?
I’ve taken other courses from Udemy that teach you how to create different neural networks (CNNs, RNNs, LSTMs, GANs, etc…) but they don’t really dive into the math behind it. This is done purposefully to get you building AI models as quickly as possible. Frameworks like PyTorch do such a great job abstracting that complexity. This course from Udacity partnering with Facebook is great because it complements the coursework I’ve done already. It really shows you the low level basics of how neural networks are created and all the math behind it. I’ll link to these courses below for anyone interested in them.