Just wanted to ask why are we not introiducing semi-supervised learning (one-class classification)
My guess is that this is just a terminology question. I had never heard the term “semi-supervised learning” before I saw your question. I googled it and it does turn out to be a more sophisticated type of machine learning that is not covered by any of the courses here, at least as far as I have seen. Here are a couple of links that talk about Semi-Supervised Learning:
But then you ask about “one class classification”. What I assume you mean by that is what is normally called “binary classification”, meaning a “yes/no” or “1/0” classification. For example, “does this picture contain a cat?”. Binary Classification is covered in MLS C1 W3: the algorithm used to do Binary Classification is called Logistic Regression and it is covered in Week 3 of Course 1.
Note that Binary Classification is a form of Supervised Machine Learning, meaning that it requires that we have “labels” for all the training data which give the correct answers for each sample (yes it is a cat or no it is not a cat).
Please let me know if I’m just missing your point here. 
“MLS” is a beginner-level course.
Semi-supervised learning is out-of-scope for the target audience.