It comes off a bit more hands on with regards to implementation. For example there was a lot of talk about efficiency and the 5th segment is a quick tutorial for Octave, since programming assignments are also part of the class. The interface of the website is better currently (and for example the slides used in the presentations are available for download). The homework section which is just review for now can be taken almost as many times as you like (the variables of the problems change ala Khan Academy) and you get instant feedback. Ng comes off a bit better at presenting his material than Norvig but not better than Thrun (this is subjective obviously). I don’t know quite what it is but there is something about the AI class which gives me the feeling that I will actually find it much more interesting than the Machine Learning class in about a month or so.
The first week of ML has proven much more demanding on my time than the AI class, but this is probably because I’ve watched most of the prerequisites for the AI class during the summer, but didn’t touch any of the stuff that’s been on the machine learning class website for two weeks or so until this weekend.
The on-line Machine Learning class seems more simplified compared to its Stanford version than the AI class is to its Stanford version (though the first week of the AI class was very easy, I understand the same lectures are used for the Stanford AI class). So far there aren’t any indications you don’t really need the AI class or something that gives equivalent knowledge before delving into the ML one.
Thanks! How does ML compare to the AI class? I considered taking both, but I was worried overextending my leisure-time learning budget.
It comes off a bit more hands on with regards to implementation. For example there was a lot of talk about efficiency and the 5th segment is a quick tutorial for Octave, since programming assignments are also part of the class. The interface of the website is better currently (and for example the slides used in the presentations are available for download). The homework section which is just review for now can be taken almost as many times as you like (the variables of the problems change ala Khan Academy) and you get instant feedback. Ng comes off a bit better at presenting his material than Norvig but not better than Thrun (this is subjective obviously). I don’t know quite what it is but there is something about the AI class which gives me the feeling that I will actually find it much more interesting than the Machine Learning class in about a month or so.
The first week of ML has proven much more demanding on my time than the AI class, but this is probably because I’ve watched most of the prerequisites for the AI class during the summer, but didn’t touch any of the stuff that’s been on the machine learning class website for two weeks or so until this weekend.
The on-line Machine Learning class seems more simplified compared to its Stanford version than the AI class is to its Stanford version (though the first week of the AI class was very easy, I understand the same lectures are used for the Stanford AI class). So far there aren’t any indications you don’t really need the AI class or something that gives equivalent knowledge before delving into the ML one.