Perhaps it's important to clarify first how complicated machine learning is. Machine learning is way more complicated than normal mathematics. The problem of adding 1+1=2 can be explained very easily, that means the description fits on a single sheet of paper and if somebody hasn't understood it, it is possible to explain it again. In contrast, machine learning is a research topic which is formulated in around 1 million academic papers and 100k full blown scientific books. And what is written in the corpus isn't completed but will extended each month by new information.
The mentioned topics of inductive learning and explanation based learning are so complicated, that the explanation needs many thousands of papers. Each of them is at least 10 pages long and additional the subject is discussed in dedicated conferences by experts. To give a general overview over the topic is possible. But not with a direct explanation but with a reference list to existing literature. If somebody has problems to understand a certain subtopic in machine learning the best idea is to provide a literature list of important milestones papers and books which can give the details. This kind of teaching situation is done usually in an academic context.
Even the famous AIMA book (Russel/Norvig) isn't able to explain directly what explanation based learning is. But they are providing an extensive literature list and if the student takes his role seriously, he has to read them all.