11–17 of 17 posts
The course material looks awesome.
The last few weeks I have been reading the first few chapters of the optional text book to get ready for the class - it all looks good, except that the programming assignments are done in Matlab or Octave. I read over the first assignment, and it will be a stretch to "love" Octave. The course material looks awesome.
I am an R to Octave convert. R has some great stuff, especially multi-type tables and a decent OO system for packaging complex analyses, but .... Octave rocks R when you have to come up with complex matrix oriented algorithms -- much, much more straightforward.
Also -- cell arrays are funky and necessary for string manipulation. I suggest not dwelling on their weirdness -- you may grow to like them.
PGM-class looks pretty hardcore. Well done Prof. Koller & Coursera team. Glad that Coursera and MITx are not watering down the material for any of their online courses.
This has been my main complaint with the Autonomous Vehicles class ran by Udacity I've done thus far. I don't think 30 minutes of video lecture per week can cover close to what an actual college class can. I ended up having to read ~2 hours on Kalman filters after completing the homework to get to a point where I actually felt like I understood the topic as well as I would in an actual class.
Was there anything aside from the derivation of the update equations for Kalman Filters you felt wasn't covered too well? What did you feel that you got from the book [and which book did you use?] that you felt wasn't covered in class?
PGM-class looks pretty hardcore. Well done Prof. Koller & Coursera team. Glad that Coursera and MITx are not watering down the material for any of their online courses.
This has been my main complaint with the Autonomous Vehicles class ran by Udacity I've done thus far. I don't think 30 minutes of video lecture per week can cover close to what an actual college class can. I ended up having to read ~2 hours on Kalman filters after completing the homework to get to a point where I actually felt like I understood the topic as well as I would in an actual class.
http://www.cs.cmu.edu/~epxing/Class/10708/lecture.html
(note the navigation bar doesn't load properly in chrome)
The last few weeks I have been reading the first few chapters of the optional text book to get ready for the class - it all looks good, except that the programming assignments are done in Matlab or Octave. I read over the first assignment, and it will be a stretch to "love" Octave. The course material looks awesome.
Re Matlab/ Octave: I am an R to Octave convert. R has some great stuff, especially multi-type tables and a decent OO system for packaging complex analyses, but .... Octave rocks R when you have to come up with complex matrix oriented algorithms -- much, much more straightforward. Also -- cell arrays are funky and necessary for string manipulation. I suggest not dwelling on their weirdness -- you may grow to like them…
There is a notable difference of course between courses like PGM and ML, which go pretty much into details of actual things, while others like AI or crypto will probably naturally stay a bit at the surface, as they are more intended to give an overview. I think they shouldn't be compared therefore. I think the robotics class is a bit behind of it's possibilities compared to PGM or ML, considering that this is definitely not an overview anymore.