The Stanford on-line AI and Machine Learning classes started last week. Andrew Ng’s machine learning class is especially interesting. I’m taking the course because there’s always a few machine learning algorithms that I haven’t actually coded up even though I know about them. For example, in my years in grad school I never once had to code up a lick of code for a neural network. NNs just aren’t a hot research topic with the profs in my department. The programming assignments are very helpful in this regard as they are very hands on.
For review purposes it’s useful to skim through the lecture videos too. For example, the first set of lectures discusses a few practical tips for debugging and parameter setting gradient descent algorithms.
Overall the course looks like it will be a lite version of the Andrew’s intro graduate course (you can find those video lectures on YouTube). It seems it will have quite a bit less theory in the homework exercises and will focus more on implementing and gaining intuition about machine learning methods. At least that is how I’m finding the first week to be.
By the way there is also a free database class that is also going on. This course seems to be an undergrad level intro into DB’s. It will cover SQL, high level DB topics and some NoSQL.