Probabilistic Graphical Models

Provided by:
8/10 stars
based on  14 reviews
Provided by:
Cost FREE
Start Date On demand

Key Concepts

lightbulb
We've created a summary of key topics covered in this course to help you decide if it's the right one for you. Click individual badges to see more courses on the same topic.

Course Details

Cost

FREE

Upcoming Schedule

  • On demand

Course Provider

Coursera online courses
Coursera's online classes are designed to help students achieve mastery over course material. Some of the best professors in the world - like neurobiology professor and author Peggy Mason from the University of Chicago, and computer science professor and Folding@Home director Vijay Pande - will supplement your knowledge through video lectures. They will also provide challenging assessments, interactive exercises during each lesson, and the opportunity to use a mobile app to keep up with yo...
Coursera's online classes are designed to help students achieve mastery over course material. Some of the best professors in the world - like neurobiology professor and author Peggy Mason from the University of Chicago, and computer science professor and Folding@Home director Vijay Pande - will supplement your knowledge through video lectures. They will also provide challenging assessments, interactive exercises during each lesson, and the opportunity to use a mobile app to keep up with your coursework. Coursera also partners with the US State Department to create “learning hubs” around the world. Students can get internet access, take courses, and participate in weekly in-person study groups to make learning even more collaborative. Begin your journey into the mysteries of the human brain by taking courses in neuroscience. Learn how to navigate the data infrastructures that multinational corporations use when you discover the world of data analysis. Follow one of Coursera’s “Skill Tracks”. Or try any one of its more than 560 available courses to help you achieve your academic and professional goals.

Provider Subject Specialization
Humanities
Sciences & Technology
4444 reviews

Course Description

In this class, you will learn the basics of the PGM representation and how to construct them, using both human knowledge and machine learning techniques.
Reviews 8/10 stars
14 Reviews for Probabilistic Graphical Models

Ratings details

  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars
  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars
  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars

Rankings are based on a provider's overall CourseTalk score, which takes into account both average rating and number of ratings. Stars round to the nearest half.

Sort By
Roman Shapovalov profile image
Roman Shapovalov profile image
10/10 starsCompleted
  • 7 reviews
  • 7 completed
4 years, 1 month ago
This course gets quite diverse reviews. Probably, some students had incorrectly estimated the difficulty and/or practical importance of the course. The course is probably one of the most challenging ones in Coursera both in terms of workload and theoretical depth. I used to spend 15+ hours per week (though it was 9-week long during the first run), and I doubt anyone could complete it successfully without basic Matlab knowledge or without prior exposure to machine learning or statistics (would be very difficult to catch up given the workload). I was a PhD student in a related field while taking the course, still the course gave me a lot of knowledge. It contains several advanced topics that are not always covered by alternative courses or books (like learning of graphical model structure). Comprehensive programming assignments (reminding small research problems) and quizzes help to make yourself really familiar with the topics. Some p... This course gets quite diverse reviews. Probably, some students had incorrectly estimated the difficulty and/or practical importance of the course. The course is probably one of the most challenging ones in Coursera both in terms of workload and theoretical depth. I used to spend 15+ hours per week (though it was 9-week long during the first run), and I doubt anyone could complete it successfully without basic Matlab knowledge or without prior exposure to machine learning or statistics (would be very difficult to catch up given the workload). I was a PhD student in a related field while taking the course, still the course gave me a lot of knowledge. It contains several advanced topics that are not always covered by alternative courses or books (like learning of graphical model structure). Comprehensive programming assignments (reminding small research problems) and quizzes help to make yourself really familiar with the topics. Some people even made T-shirts with a caption “I survived PA5” — that was an assignment on Gibbs sampling, was challenging indeed, but the experience was that rewarding. Some PA’s really require figuring out the data structures involved, but this is a compromise between having students developing complex software from scratch and spending reasonable amount of time. Daphne, as Coursera co-founder, made her best to show the capabilities of the platform (it was one of the first MOOCs). Unlike some other courses, you feel like you are really taking a quality graduate- level course comparable (though not identical) to the ones the author teaches at her university. To sum up, prospective students should take into account that the course is quite advanced (vs. introductory), academic (vs. practical), and demanding (vs. laid-back), when deciding whether it fits their goals.
Was this review helpful? Yes6
 Flag
Student profile image
Student profile image

Student

10/10 starsTaking Now
2 years, 8 months ago
Course contents are highly standard and topics are very useful and advanced but the course prerequisite very good knowledge on these topics. Assignments standard are not good. They should focus on basic learning and more intuitive.
Was this review helpful? Yes0
 Flag
Dan Friedman profile image
Dan Friedman profile image
10/10 starsTaking Now
  • 1 review
  • 0 completed
3 years, 4 months ago
I am very satisfied with this course - I am half way through it, and I feel like it gives me a thorough understanding of the building blocks of Probabilistic Graphical Models. After finishing week 4, I can safely say that I know what are Bayesian Networks and Markov Random Fields, how and when to use them, and why they work. I have a feeling that the course is taught very clearly - What once seemed to me like difficult concepts suddenly seem actually fairly simple - which is definitely a result of the quality of the teaching. This is NOT an easy course by all means - Each week has around 3 hours of lectures which are not very easy - It takes me about 5 hours to complete them. There is a quiz after each lecture which takes me around 30 minutes to complete, and the programming exercises require you implement fully the data structures and algorithms. For me this means that each week I need to spend around 10 hours on the programming exe... I am very satisfied with this course - I am half way through it, and I feel like it gives me a thorough understanding of the building blocks of Probabilistic Graphical Models. After finishing week 4, I can safely say that I know what are Bayesian Networks and Markov Random Fields, how and when to use them, and why they work. I have a feeling that the course is taught very clearly - What once seemed to me like difficult concepts suddenly seem actually fairly simple - which is definitely a result of the quality of the teaching. This is NOT an easy course by all means - Each week has around 3 hours of lectures which are not very easy - It takes me about 5 hours to complete them. There is a quiz after each lecture which takes me around 30 minutes to complete, and the programming exercises require you implement fully the data structures and algorithms. For me this means that each week I need to spend around 10 hours on the programming exercises, so overall I would say the course requires ~15 hours of work each week, for someone with a similar background as me. I would imagine that for people who are new to computer science or machine learning, this would probably be a VERY challenging course - It does not make any discounts since its an online course, and I would say it is a graduate level Stanford course. My background: I have an Msc. in Computational Biology, and some programming experience in Matlab and R. I have a relatively sound statistical understanding, and this is not the first time I deal with graphical models, but is the the first time I study it in depth. I am doing the course 2 years after it was first run - It is too bad that you can't join it at any given point - it would have been nice to get a Coursera certificate for completing the course, as right now I get my exercises graded, but then I'm told that I don't get an official grade since I'm past the deadline... Though as I'm in it to learn, it doesn't really matter to me. Good luck!
Was this review helpful? Yes0
 Flag
username profile image
username profile image
7/10 starsDropped
  • 3 reviews
  • 0 completed
3 years, 11 months ago
I had to drop this one due to its difficulty. I wrote had free time, watched the lectures, wrote things down, but when the Bayesian network homework came up, I couldn't get the numbers to check out.
Was this review helpful? Yes1
 Flag
Rafael Santos profile image
Rafael Santos profile image
5/10 starsCompleted
  • 4 reviews
  • 4 completed
4 years ago
This course was extremely hard and very time consuming. I followed until the end but naturally did not get the minimum points to get the certificate as I did not have the time to study properly. But I'll still do it again once I have a more free agenda. So if you have a busy week, work or college, you better consider if this is the appropriate time. About the course, I think it is a very good opportunity to have a course taught at Stanford level available to everyone. Most of the courses on Coursera are introduction courses and PGM is the kind of course that comes up to defy a avid for challenge student. It has a lot of math and demands many requirements specially from Statistics. One thing I disliked in the course is the lack of more practical examples. Also the strict deadlines.
Was this review helpful? Yes0
 Flag
Anna Nachesa profile image
Anna Nachesa profile image
10/10 starsCompleted
  • 13 reviews
  • 13 completed
4 years, 6 months ago
One of the most technically challenging, but also one of the most interesting courses I have taken. Loved the subject, hated that I could not spent more time on it. The book (which I bought) helps a lot, if you are serious about the subject, I recommend getting it. This is one of the subjects where the lectures aren't enough, they can at best serve as a guideline and the study effort is yours! Regarding the prerequisites: some knowledge of probability theory and of programming is a must. Previous exposure to Octave/Matlab might help too.
Was this review helpful? Yes4
 Flag
Student profile image
Student profile image

Student

10/10 starsCompleted
4 years, 5 months ago
Although it is a very(!) demanding course with a difficult subject to learn, I found this course to be very rewarding. In the past I've learnt this subject from other sources and it was a mess. I think its the best course I've taken in coursera till now, and I strongly recommend it for getting a strong basis in PGM.
Was this review helpful? Yes0
 Flag
Equanimous Creativity profile image
Equanimous Creativity profile image
10/10 starsCompleted
  • 33 reviews
  • 32 completed
4 years, 7 months ago
This course is very hard and very useful. Knowing how to build Probabilistic Graphical Models is fundemental if you need to build models from data.
Was this review helpful? Yes0
 Flag
George Terzakis profile image
George Terzakis profile image
10/10 starsCompleted
  • 1 review
  • 1 completed
4 years, 9 months ago
This course has been by far the most helpful course in Coursera for me. Many people can argue that it is a bit hard, it requires somewhat advanced skills in probability and possibly a few other topics in mathematics; I think really it is nothing extreme. Anyway, hard-difficult is exactly what the material is about. Most importantly, for some of us who, for a variety of reasons- don't get to have specialized (in the field of AI) college instructors like Daphne Koller, the chance to take a course more-less the way it's given at Stanford is highly appreciated (to say the least). I am a graduate student and my background already includes calculus, probabilities, etc, so yet another introduction wouldn't do me any good. On the other hand, "Probabilistic Graphical Models" is a modern AI approach and the concepts are very difficult to read from a book alone (mainly because of the -somewhat inefficient for learning- ways of illustrating grap... This course has been by far the most helpful course in Coursera for me. Many people can argue that it is a bit hard, it requires somewhat advanced skills in probability and possibly a few other topics in mathematics; I think really it is nothing extreme. Anyway, hard-difficult is exactly what the material is about. Most importantly, for some of us who, for a variety of reasons- don't get to have specialized (in the field of AI) college instructors like Daphne Koller, the chance to take a course more-less the way it's given at Stanford is highly appreciated (to say the least). I am a graduate student and my background already includes calculus, probabilities, etc, so yet another introduction wouldn't do me any good. On the other hand, "Probabilistic Graphical Models" is a modern AI approach and the concepts are very difficult to read from a book alone (mainly because of the -somewhat inefficient for learning- ways of illustrating graph structures with mathematical formulas). Moreover, the course is not exactly found in every graduate program in existence. So, having the ability to take such a course was literally a blessing, especially for my work and I am very thankful for this. I think people should not get frustrated with the material. It really is somewhat advanced, but that is how it should be. A simplified version is not going to be of any use to anyone really. As far as I have seen, Coursera tries to cover a wide range of backgrounds, ages, languages, etc., so it makes sense that some courses are more difficult than others.
Was this review helpful? Yes2
 Flag
emediquei profile image
emediquei profile image
8/10 starsCompleted
  • 5 reviews
  • 5 completed
4 years, 9 months ago
Definitely a challenging course even for someone with some university level mathematics and computer science background.
Was this review helpful? Yes0
 Flag
Thomas Johnson profile image
Thomas Johnson profile image
3/10 starsCompleted
  • 3 reviews
  • 3 completed
4 years, 11 months ago
Lots of math, but few examples of how to apply it. Homework assignments are terrible - more than half the time is spent figuring out the intricacies of the ad-hoc data structures used rather than applying the techniques learned in class.
Was this review helpful? Yes2
 Flag
Katie Cornog profile image
Katie Cornog profile image
8/10 starsCompleted
  • 1 review
  • 1 completed
4 years, 11 months ago
This course covers a lot of material. Although it is advertised as taking 8-10 hours per week, it takes almost everyone closer to 15-20. This course is taught as a Stanford graduate course, and is considered a hard one even at Stanford. To get the most out of the course you should have a fairly good knowledge of probability and statistics. For example, do you know clearly what a marginal distribution is? a conditional distribution? Are you very comfortable with Bayes rule? You should also be comfortable programming in Matlab. There are 9 programming assignments. A structured outline of code is given, but you must be able to understand it an fill in the missing algorithmic parts. The code can be confusing to interpret due to the use of many advanced Matlab features. The programming assignments were the best aspect of the course for me because being guided through an implementation of the techniques makes them much clearer than reading... This course covers a lot of material. Although it is advertised as taking 8-10 hours per week, it takes almost everyone closer to 15-20. This course is taught as a Stanford graduate course, and is considered a hard one even at Stanford. To get the most out of the course you should have a fairly good knowledge of probability and statistics. For example, do you know clearly what a marginal distribution is? a conditional distribution? Are you very comfortable with Bayes rule? You should also be comfortable programming in Matlab. There are 9 programming assignments. A structured outline of code is given, but you must be able to understand it an fill in the missing algorithmic parts. The code can be confusing to interpret due to the use of many advanced Matlab features. The programming assignments were the best aspect of the course for me because being guided through an implementation of the techniques makes them much clearer than reading a book or listening to a lecture. There is an active online community for this course. This is helpful for understanding some of the ambiguities in the lectures and homework assignments. I found the textbook, although not required, to be a big help in trying to do the homework. Often topics were only touched on superficially in the lectures and yet the homework questions were very detailed. The book filled in the needed material.
Was this review helpful? Yes1
 Flag
Kaushalya Madhawa profile image
Kaushalya Madhawa profile image
7/10 starsCompleted
  • 3 reviews
  • 3 completed
3 years, 10 months ago
This is one of the hardest courses I've taken. Due to lack of concrete examples some concepts are hard to grasp just by watching the videos. If a high-level overview can be given before introduction of a new concept it would be easier to understand new equations/concepts. Prof.Yaser S. Abu-Mostafa does this in his "Learning From data" Course.
Was this review helpful? Yes1
 Flag
Sai profile image
Sai profile image
8/10 starsCompleted
  • 13 reviews
  • 13 completed
4 years, 12 months ago
This is a graduate level course on a branch of machine learning taught by a co-founder of Coursera. This is one of the toughest course that I have taken so far. Prof. Koller admits that this is a tough one for Stanford students as well. The course page indicates a workload of 8-10 hours per week but expect to spend twice as much. The biggest difficulty I faced with was that the instructions of the programming assignments were poorly written and we had to consult the forum a lot to figure out what we were expected to do. The situation was gradually improved toward the end of the course, though. Please note that I took the first offering of this course in March 2012 and things might be different by now. All programming assignments are in Octave (or Matlab if you can afford it). There were 2 tracks: basic and advanced. Programming assignments were required only for the advanced track. Here is some statistics that I found in the forum: 4... This is a graduate level course on a branch of machine learning taught by a co-founder of Coursera. This is one of the toughest course that I have taken so far. Prof. Koller admits that this is a tough one for Stanford students as well. The course page indicates a workload of 8-10 hours per week but expect to spend twice as much. The biggest difficulty I faced with was that the instructions of the programming assignments were poorly written and we had to consult the forum a lot to figure out what we were expected to do. The situation was gradually improved toward the end of the course, though. Please note that I took the first offering of this course in March 2012 and things might be different by now. All programming assignments are in Octave (or Matlab if you can afford it). There were 2 tracks: basic and advanced. Programming assignments were required only for the advanced track. Here is some statistics that I found in the forum: 44,000 students registered, 6,450 students attempted week 1 quiz, 3,070 week 2 quiz, and about 1,100 finished the last programming assignment. Not directly related to this course but you may find Prof. Koller's TED talk interesting: http://www.ted.com/talks/daphne_ koller_what_we_re_learning_from_online_education.html
Was this review helpful? Yes1
 Flag

Rating Details


  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars
  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars
  • 5 stars
  • 4 stars
  • 3 stars
  • 2 stars
  • 1 stars

Rankings are based on a provider's overall CourseTalk score, which takes into account both average rating and number of ratings. Stars round to the nearest half.