Of course I am grateful for all the people who put in effort to teach people things, and then doubly grateful for people who make it accessible and triply grateful for those who make it free.
At this point a guide to machine learning guides would also be a huge contribution. Some ideas for ranking/filtering/grading features: how much math, how much hands on, which language, assumed knowledge and pedagogical strategy/path.
Seriously, how do these kinds of things keep hitting first page HN? Quality aside, who on HN thinks this is of interest to the community in terms of being novel / worth discussing?
i'm gonna get downvoted for elitism but it's very clearly "aspirational upvotes" from people that don't actually know any ML and won't ever end up learning any. to wit: no one that actually studies seriously spends this much time wringing their hands over which reference to use.
you don't keep cycling through references that seem better because that's a sure way to never make any progress - you pick a reference and grind through it. maybe with occasional double checking against some other reference sure but you never end up ditching the first one because the cost of changing notations/formalisms/etc is very very high and not worth paying almost ever (since intrinsically they're all talking about the same thing anyway). this is the reason that profs today still teach from the books they learned from 30 years ago.
The intro seems really funny to me. They could find no textbook that covers these topics so they printed their own. Meanwhile ISLR just released their 2nd edition. This book is the gold standard of all textbooks let alone on statistical learning. Their efforts would have been better covering case studies or advanced applications well.
Happy to see that at least the intro is being read by soneone ;) Yes, ISLR is a great book, but it's intended for a different audience. When we started teaching ML, we did use ISLR but we weren't happy with it because it was too light on the math side. When introducing ML to people who already has taken some math courses, it felt pity not to leverage their calculus and linear algebra knowledge - ISLR hardly mention vectors and derivatives... (and, to be honest... who wants to use R? But that wasn't any major reason) So we put together our own material for that course. And it felt worth the extra effort to turn it into a textbook available to anyone on the internet, and not just a pdf looked up somewhere in a university "learning platform" accessible only for a bunch of selected students. Maybe that wasn't the right priority, I don't know, but that's what we did.
7 comments
[ 3.2 ms ] story [ 25.6 ms ] threadAt this point a guide to machine learning guides would also be a huge contribution. Some ideas for ranking/filtering/grading features: how much math, how much hands on, which language, assumed knowledge and pedagogical strategy/path.
you don't keep cycling through references that seem better because that's a sure way to never make any progress - you pick a reference and grind through it. maybe with occasional double checking against some other reference sure but you never end up ditching the first one because the cost of changing notations/formalisms/etc is very very high and not worth paying almost ever (since intrinsically they're all talking about the same thing anyway). this is the reason that profs today still teach from the books they learned from 30 years ago.
/Andreas, the first author