Knowledge engineering and how it is complementary to Deep Learning

I was reading this stackoverflow’s blog. He is talking about rules being learnt. I am a bit lost trying to understand what he seems to be talking about. Does anyone knows any good articles that can help me get some context on how this stuff works?

This blog post contains link to this article which essentially seems like the same thinking that Jeremy has been talking about in these courses. This is unrelated to the question above. Just thought might be interesting read.