I just recently finished a Coursera online learning course of Machine Learning. The course I chose was Andrew Ng’s Stanford University course “Machine Learning”. Having an Electrical Engineering background pays off, the discussions of partial derivatives is something an EE should be comfortable with.
The course is well laid out and well structured, though a little light on the programming side. The course relies on Octave, an opensource clone of MatLab, for all programming assignments. The assignments for the most part are very straight forward, the focus is on the concepts (“intuition” as it is commonly referred to throughout the course) and not any advance programming techniques.
What is apparent is that for most ML/DL projects, the actual implementation of the ML approach is something handled by tried and true libraries. This doesn’t make an ML project simple, rather it moves the focus to the process of defining the problem and finding the optimal solution.
So now that I have the fundamentals of ML in my toolbag it’s time to create an ML solution. My goal is to do this all in Python. I have some projects in mind, once I get started I’ll make sure I update my status here.