I'm in the middle of the machine learning coursera course, and registered for this one as well due to interest in the material.<p>My one complaint is that the programming assignments weren't interesting at all. The results were interesting, but the setups were mostly given to us, and we just had to code an algorithm that was in our notes. For someone who understands the basics of linear algebra and programming, it was just a syntax challenge, and that got irritating after a bit so I stopped doing them.<p>I won't get the certificate for completing the course, but I have a few extra hours of free time each week to add this second course, so I'm happy. I doubt that the actual homework that Stanford students taking this course get is so easy and repetitive, though, and I'm positive they wouldn't complain about not getting to retake quizzes after getting poor grades.<p>Not to knock the course. I've learned a lot and the professor (Andrew Ng) does a good job.
Hinton is a huge figure in the neural network literature and an important researcher in deep learning. After going through the first week of lectures, I can say he's also an excellent teacher.<p>The syllabus, draft though it is, indicates the second half of the class will focus on deep learning, a field of machine learning that has demonstrated huge potential.
Just browsing through the Coursera Computer Science listings, it looks like they are rapidly approaching the point where you could put together a CS curriculum superior to what you could get at any single school. The people they have teaching a lot of these topics are some of the best in the world in their field. The Micahel Collins NLP course looks really thorough and up to date, for example I took a similar course a few years ago, and I remember reading papers written by him.<p>As has been said by many already, of course, the remaining nuts to crack are high quality interaction with other students, professors, and TAs; and accreditation.<p>But the dis-intermediation of large universities may be nearer than we think.
people are already complaining that you can only take the quizes once ... he had to send out an email today to everyone saying:<p>"Many of you are unhappy with only being allowed to attempt a quiz once. Starting in week two, we have therefore decided to make up twice as many questions and to allow you to do each quiz twice if you want to. The second time you try it the questions will all be different. Your score will be the maximum of your two scores. For week one, the quizzes will remain as they are now.<p>Many of you would like the names of the videos to be more informative. We will change the names to indicate the content and the duration.<p>Some of you thought that some of the quiz questions were too vague. We will try to make future questions less vague.<p>Some of you are unhappy that we do not have the resources to support Python for the programming assignments. We sympathize with you and would do it if we could. You are still welcome to use Python (or any other language) if you can port the octave starter code to your preferred language. We have no objection to people sharing the ported versions of the starter code (but only the starter code!). However, if you get starter code in another language from someone else, you are responsible for making sure it does not contain bugs."<p>I thought that was pretty funny!
The only course that is not significantly diluted is Koller's PGM. All others have been dumbed down to a degree where they provide no challenge to the courseree at all.
> Neural Networks are gradually taking over from simpler Machine Learning methods<p>And haven't SVMs and such gradually taken over from Neural Networks?
I tried to do a couple coursera courses and found the video lectures highly inefficient; very needlessly time consuming, even watching them sped up. All I really want is a glorified text book with quiz grading and a final.