7 ms·

Complete Course on Machine Learning

268 points by weeha 11y ago

gamapuna 11y ago
Here's the complete course: http://alex.smola.org/teaching/cmu2013-10-701/ http://alex.smola.org/teaching/cmu2013-10-701/
- ojaved 11y ago
  This is the 2013 course. The original post is for the spring 2015 machine learning course
- bcheung 11y ago
  The videos don't seem to work.
joshvm 11y ago
Some nice courses there, also check out Dan Cremer's lectures on variational methods for computer vision if you're interested in that sort of thing. There's also a nice series on computer vision for special effects. http://www.computervisiontalks.com/variational-methods-for-computer-vision-lecture-2-prof-daniel-cremers/ http://www.computervisiontalks.com/variational-methods-for-c...
zablocky 11y ago
Have anyone seen some and can tell if the material is well explained?
- smilekzs 11y ago
  I took the course (Alex Smola's 10-701, Spring '15) in the classroom. Personally I don't like his lecture style -- too vague, too many assumptions, too much reliance on jargon he hasn't already explained. YMMV.
rugatelstvo 11y ago
I am under the impression that to learn statistics one must first have a working knowledge of probability theory which rests upon grad level math analysis. Can machine learning be studied without any of that?
- solomatov 11y ago
  I think, you can get away without mathematical analysis (however, it's not that complex). However, basic probability theory is a must have.
- pvnick 11y ago
  What do you mean, grad level math analysis? Much of probability theory can be learned with basic multivariate calculus. (Perhaps there's a terminology misunderstanding here - when I see "grad level" I think "grad school," ie masters/phd). Certainly basic probability theory is a plus.
  - rugatelstvo 11y ago
    By grad level analysis I mean analysis based on Measure Theory. Here's where I got the idea(last comment in the linked thread): https://www.physicsforums.com/threads/what-is-the-most-useful-math.187372/ https://www.physicsforums.com/threads/what-is-the-most-usefu...
    - huac 11y ago
      Measure theory being necessary to statistics is rather contentious; a better discussion is on Andrew Gelman's blog [1]. My school's PhD stats program does require real analysis before the prelims, but for most intents and purposes, 'multi' and 'linal' (as the cool kids say) should be sufficient for machine learning from a comp sci perspective. I haven't fully worked through ESLR (Hastie and Tibsharini's advanced version of ISLR posted above) but the majority of the math there is linear algebra with some differential equations and calculus thrown in. I've heard Harvard Stat 210 and Berkeley Stat 205A/B cited as good examples of mathematical stat classes - if you're seriously interested maybe take a look at those syllabi. [1]: http://andrewgelman.com/2008/01/14/what_to_learn_i/ http://andrewgelman.com/2008/01/14/what_to_learn_i/
anacleto 11y ago
That's really nice. Dan Cremer is impressive. Here's a great Laboratory on Amazon ML for Human Activity Recognition (w/ Python). https://cloudacademy.com/amazon-web-services/labs/aws-machine-learning-human-activity-recognition-21/ https://cloudacademy.com/amazon-web-services/labs/aws-machin... Totally worth a look.
btown 11y ago
I like his teaching style, but it seems some of the lecture videos (1.3, for example) are cut off - very frustrating! For anyone watching nonetheless, I recommend going into YouTube and changing the speed to 1.5x.
- ojaved 11y ago
  The lectures on computervisiontalks are directly being taken from youtube (but tags, navigation, bookmarking and in-video search capability is added). The lecture 1.3 (for spring 2015 class) is exactly of the same length. However on youtube, the lectures for machine learning class 2013 (also by Alex Smola) are available which are of a different length.
yla92 11y ago
A bit off topic : what are the best recommended way/resources to learn linear algebra and basic probability and statistics ?
- smockman36 11y ago
  I think this is a good resource to start: http://betterexplained.com/articles/linear-algebra-guide/ http://betterexplained.com/articles/linear-algebra-guide/
- id_ris 11y ago
  For linear algebra check out Prof. Gilbert Strand's course on MIT OCW. He's great at explaining the material and the course resources are comprehensive. http://ocw.mit.edu/courses/mathematics/18-06-linear-algebra-spring-2010/video-lectures/ http://ocw.mit.edu/courses/mathematics/18-06-linear-algebra-...
- stdbrouw 11y ago
  For linear algebra, I like the "No BS guide to linear algebra" (https://gumroad.com/l/noBSLA https://gumroad.com/l/noBSLA) which also includes a high school math refresher for people who need it (I did). For probability, "Probability Demystified" is a good basic intro. For statistics, I would really recommend Allen Downey's Think Stats (http://greenteapress.com/thinkstats2/index.html http://greenteapress.com/thinkstats2/index.html), especially if you're coming from a programming background. Most introductions to statistics focus heavily on the mathematics needed to enable certain analytical approximations to difficult probabilistic calculations (e.g. the t-test), whereas Think Stats just bites that bullet and focuses on simulation / brute force so you can spend more time on the actual fundamental theory behind statistics. Brian Blais' "Statistical Inference for Everyone" (http://web.bryant.edu/~bblais/statistical-inference-for-everyone-sie.html http://web.bryant.edu/~bblais/statistical-inference-for-ever...) also looks really good, but haven't had a chance to review it in depth.
- delluminatus 11y ago
  Depends on how basic you're imagining. Khan Academy [0] is a fairly well-regarded free resource for high-school and undergraduate level mathematics video lectures. They have probability and statistics as well as linear algebra courses. If you prefer textbooks, I have heard good things about "Linear Algebra Done Right," [1] but I would not recommend it unless you are "math literate" at an undergraduate level already. [0] https://www.khanacademy.org/math/probability https://www.khanacademy.org/math/probability [1] http://www.amazon.com/dp/3319110799 http://www.amazon.com/dp/3319110799
chrisdbaldwin 11y ago
Some of the videos in the link are cut short, and the full videos are much better. Here's a link to a playlist of the full lectures: https://www.youtube.com/playlist?list=PLZSO_6-bSqHQmMKwWVvYwKreGu4b4kMU9 https://www.youtube.com/playlist?list=PLZSO_6-bSqHQmMKwWVvYw...
- ojaved 11y ago
  The videos on computervisiontalks.com are exactly the same as videos on youtube because the site is pulling these videos from youtube. The post points to the spring 2015 lectures. You are pointing to earlier lectures in 2014,2013
ericmo 11y ago
I like Smola's ML book, and it's great to see a full-depth ML course online, I'll certainly watch some videos. Other than that, the audio quality could be better.
phunehehe0 11y ago
Just want to shout about this very comprehensive course by Caltech professor Yaser S. Abu-Mostafa http://work.caltech.edu/telecourse.html http://work.caltech.edu/telecourse.html

268 points by weeha 11y ago

Here's the complete course: http://alex.smola.org/teaching/cmu2013-10-701/ http://alex.smola.org/teaching/cmu2013-10-701/

ojaved 11y ago
This is the 2013 course. The original post is for the spring 2015 machine learning course
bcheung 11y ago
The videos don't seem to work.

joshvm 11y ago

Some nice courses there, also check out Dan Cremer's lectures on variational methods for computer vision if you're interested in that sort of thing. There's also a nice series on computer vision for special effects. http://www.computervisiontalks.com/variational-methods-for-computer-vision-lecture-2-prof-daniel-cremers/ http://www.computervisiontalks.com/variational-methods-for-c...

zablocky 11y ago

Have anyone seen some and can tell if the material is well explained?

smilekzs 11y ago
I took the course (Alex Smola's 10-701, Spring '15) in the classroom. Personally I don't like his lecture style -- too vague, too many assumptions, too much reliance on jargon he hasn't already explained. YMMV.

rugatelstvo 11y ago

I am under the impression that to learn statistics one must first have a working knowledge of probability theory which rests upon grad level math analysis. Can machine learning be studied without any of that?

solomatov 11y ago
I think, you can get away without mathematical analysis (however, it's not that complex). However, basic probability theory is a must have.
pvnick 11y ago
What do you mean, grad level math analysis? Much of probability theory can be learned with basic multivariate calculus. (Perhaps there's a terminology misunderstanding here - when I see "grad level" I think "grad school," ie masters/phd). Certainly basic probability theory is a plus.
- rugatelstvo 11y ago
  By grad level analysis I mean analysis based on Measure Theory. Here's where I got the idea(last comment in the linked thread): https://www.physicsforums.com/threads/what-is-the-most-useful-math.187372/ https://www.physicsforums.com/threads/what-is-the-most-usefu...
  - huac 11y ago
    Measure theory being necessary to statistics is rather contentious; a better discussion is on Andrew Gelman's blog [1]. My school's PhD stats program does require real analysis before the prelims, but for most intents and purposes, 'multi' and 'linal' (as the cool kids say) should be sufficient for machine learning from a comp sci perspective. I haven't fully worked through ESLR (Hastie and Tibsharini's advanced version of ISLR posted above) but the majority of the math there is linear algebra with some differential equations and calculus thrown in. I've heard Harvard Stat 210 and Berkeley Stat 205A/B cited as good examples of mathematical stat classes - if you're seriously interested maybe take a look at those syllabi. [1]: http://andrewgelman.com/2008/01/14/what_to_learn_i/ http://andrewgelman.com/2008/01/14/what_to_learn_i/

anacleto 11y ago

That's really nice. Dan Cremer is impressive. Here's a great Laboratory on Amazon ML for Human Activity Recognition (w/ Python). https://cloudacademy.com/amazon-web-services/labs/aws-machine-learning-human-activity-recognition-21/ https://cloudacademy.com/amazon-web-services/labs/aws-machin... Totally worth a look.

btown 11y ago

I like his teaching style, but it seems some of the lecture videos (1.3, for example) are cut off - very frustrating! For anyone watching nonetheless, I recommend going into YouTube and changing the speed to 1.5x.

ojaved 11y ago
The lectures on computervisiontalks are directly being taken from youtube (but tags, navigation, bookmarking and in-video search capability is added). The lecture 1.3 (for spring 2015 class) is exactly of the same length. However on youtube, the lectures for machine learning class 2013 (also by Alex Smola) are available which are of a different length.

yla92 11y ago

A bit off topic : what are the best recommended way/resources to learn linear algebra and basic probability and statistics ?

smockman36 11y ago
I think this is a good resource to start: http://betterexplained.com/articles/linear-algebra-guide/ http://betterexplained.com/articles/linear-algebra-guide/
id_ris 11y ago
For linear algebra check out Prof. Gilbert Strand's course on MIT OCW. He's great at explaining the material and the course resources are comprehensive. http://ocw.mit.edu/courses/mathematics/18-06-linear-algebra-spring-2010/video-lectures/ http://ocw.mit.edu/courses/mathematics/18-06-linear-algebra-...
stdbrouw 11y ago
For linear algebra, I like the "No BS guide to linear algebra" (https://gumroad.com/l/noBSLA https://gumroad.com/l/noBSLA) which also includes a high school math refresher for people who need it (I did). For probability, "Probability Demystified" is a good basic intro. For statistics, I would really recommend Allen Downey's Think Stats (http://greenteapress.com/thinkstats2/index.html http://greenteapress.com/thinkstats2/index.html), especially if you're coming from a programming background. Most introductions to statistics focus heavily on the mathematics needed to enable certain analytical approximations to difficult probabilistic calculations (e.g. the t-test), whereas Think Stats just bites that bullet and focuses on simulation / brute force so you can spend more time on the actual fundamental theory behind statistics. Brian Blais' "Statistical Inference for Everyone" (http://web.bryant.edu/~bblais/statistical-inference-for-everyone-sie.html http://web.bryant.edu/~bblais/statistical-inference-for-ever...) also looks really good, but haven't had a chance to review it in depth.
delluminatus 11y ago
Depends on how basic you're imagining. Khan Academy [0] is a fairly well-regarded free resource for high-school and undergraduate level mathematics video lectures. They have probability and statistics as well as linear algebra courses. If you prefer textbooks, I have heard good things about "Linear Algebra Done Right," [1] but I would not recommend it unless you are "math literate" at an undergraduate level already. [0] https://www.khanacademy.org/math/probability https://www.khanacademy.org/math/probability [1] http://www.amazon.com/dp/3319110799 http://www.amazon.com/dp/3319110799

chrisdbaldwin 11y ago

Some of the videos in the link are cut short, and the full videos are much better. Here's a link to a playlist of the full lectures: https://www.youtube.com/playlist?list=PLZSO_6-bSqHQmMKwWVvYwKreGu4b4kMU9 https://www.youtube.com/playlist?list=PLZSO_6-bSqHQmMKwWVvYw...

ojaved 11y ago
The videos on computervisiontalks.com are exactly the same as videos on youtube because the site is pulling these videos from youtube. The post points to the spring 2015 lectures. You are pointing to earlier lectures in 2014,2013

ericmo 11y ago

I like Smola's ML book, and it's great to see a full-depth ML course online, I'll certainly watch some videos. Other than that, the audio quality could be better.

phunehehe0 11y ago

Just want to shout about this very comprehensive course by Caltech professor Yaser S. Abu-Mostafa http://work.caltech.edu/telecourse.html http://work.caltech.edu/telecourse.html