I don’t expect everyone to agree with me but I think it’s important to voice all opinions so that they be taken into consideration when further improving the course. I also invite others to pitch in. But before going into it, I’d like to say that I also took a quick look at the old version of the course before voicing my opinion.
I feel like the course has been dumbed down to a fault. I do understand that this decision was probably made in order to make it accessible to a wider audience but some of the changes seem detrimental and, at times, perhaps a bit… insulting. Just to give a couple of examples:
-
The pace is terribly slow. I finished the first two weeks in much less than a day and would have probably finished all three if it weren’t for the fact that it’s really hard to not get distracted when there is essentially nothing being said for half a minute straight every other minute. I understand that different people have different paces, that they have different backgrounds, and a different amount of spare time but it’s not like they need to be told that they’re on a field and looking all around before going down the biggest slope 4 times before going “Aha! Now I get it! The fourth step really locked it in for me.”
-
The non-standard terminology seems to cause more harm than good to me. In order to avoid potential slight confusion that might last for a couple of seconds, learners end up not becoming familiarized with standard terms that they might encounter in the future. For example, it is only briefly mentioned that model’s function is called its hypothesis and afterwards it’s simply referred to as “function.” Why is that? It’s not like people would become be so confused by this term that their minds would explode. It’s not even such an unreasonable term, given that we can think of the model being improved by gradient descent making better and better hypotheses about it. Similarly, why is theta not used any longer? It’s the standard notation used pretty much everywhere.
-
Who is your target audience? Is this supposed to be both an introduction to machine learning and to programming? Because I expect mostly no one falls under that category.
That said, I do appreciate all the hard work that has been put into developing this course. Everything said above is only meant as constructive criticism given with much love. It’s just that I’m not the best at making it sound less harsh.
Cheers,
Bogdan