I took this course when it was first offered in 2012 and I can't recommend it highly enough. It has had a big impact on my career.
The course focuses almost entirely on neural nets rather than taking detours through the rest of machine learning. I recently learned that the well-known optimization technique RMSProp, presented in the slides of this course[1], was not previously published and subsequent papers have been using the course notes as a citation for it. Despite this level of cutting edge knowledge being contained in the course, it was possible (though difficult) for me to follow with no previous machine learning experience and no prerequisites other than the math I learned in the course of getting a CS degree.
Pretty frustrated about this course. I had registered for the 2012 version at the end of June 2016 and after I got a quarter of a way through the lessons they removed it. When I talked to the company they said it would be re-available in september. I'm skeptical it will contain new material and am extremely frustrated with the overall experience.
They seem to have done a purge on older courses recently, lost access to 2 courses recently as well. It's extremely frustrating to adapt to the particular style and flow a course and then have it ripped out for...? (I'm not actually sure what justification goes here)
Unless I can download all of the content at once, I'm staying away. (This works well for textbooks, but videos can be large. ~40GB/Semester/Class for 720p MIT OCW off youtube)
4 comments
[ 2.7 ms ] story [ 20.6 ms ] threadThe course focuses almost entirely on neural nets rather than taking detours through the rest of machine learning. I recently learned that the well-known optimization technique RMSProp, presented in the slides of this course[1], was not previously published and subsequent papers have been using the course notes as a citation for it. Despite this level of cutting edge knowledge being contained in the course, it was possible (though difficult) for me to follow with no previous machine learning experience and no prerequisites other than the math I learned in the course of getting a CS degree.
[1] http://www.cs.toronto.edu/~tijmen/csc321/slides/lecture_slid...
Unless I can download all of the content at once, I'm staying away. (This works well for textbooks, but videos can be large. ~40GB/Semester/Class for 720p MIT OCW off youtube)