By P. Lepin 6 months ago Completed

An excellent introduction into essential machine learning techniques. The course is very rich in content, and covers a lot of ground, but doesn't ever devolve into empty hand-waving. The course favours practical approach to machine learning, and will often skip the theory and/or underlying principles (leaving formula derivation as a purely optional exercise for those interested in this aspect of ML).

Prof. Ng is obviously enthusiastic about the subject, and the course as a whole feels very polished.

On the downside, the programming assignments are not very challenging and do not require any creativity, as they boil down to following very detailed instructions. The assignments remain quite instructive despite that, as there's a lot of support code meant to visualize the results and provide various statistics to help students understand how does everything work. This doesn't seem to be an oversight or anything like that, but rather conscious course design as a 'ML cookbook'.

Since going through this class last spring I actually employed a few of the techniques taught in my day-to-day work, and this class was instrumental in sparkling my newfound interest for statistics.

Required skills: elementary algebra, coding skills

Recommended skills: first-order logic, linear algebra, probability & statistics, multivariate calculus, Octave

Workload: low
Difficulty: low
Value: high
Fun: high

By Gavin Conran from Portadown 7 months ago Completed

This was the first MOOC I took and have completed a number of them since. Some have been wonderful but Andrew's ML course still reigns supreme.

By Alex Parij from Montreal, Quebec 4 months ago Completed

Overall the course was interesting.
I wish the programming assignments were more engaging and not just to fill in couple of lines in Octave code.

By Sebastián Ramírez Montaño from Bogota, Bogota D.C. 5 months ago Completed

One of the greatest!

Professor Andrew Ng is great, he makes you understand and doesn't try to make you feel dumb, he explains it all that you need to use Machine Learning without overwhelming you with mathematical complexities.

This course has many of the greatest Machine Learning algorithms that you can use to work in many many applications.

By Panagiotis from New York, New York 5 months ago Completed

Great introductory course! The discussion forum is really useful.. The programming assignments are meaningful and to the point..

By Daniel Snider from Toronto, Ontario 5 months ago Completed

I loved tis class so much. It was brilliantly taught which made for an extremely enjoyable learning process. This course makes the Coursera system rock.

By Afref Fetter 5 months ago Completed

Prior experience in the field: None

Like:
We are introduced to a wide array of topics from basic regression to SVMs.

Practical applications of the techniques was shown in large-scale projects.

We got to implement what we'd learned in the lectures through some excellent (and useful) programming assignments.

Dislike:
The course left me feeling I had only an "overview" of machine learning, rather than being able to say I'd learned the nitty-gritty details [This could be a good thing depending on what you want].

The quizzes didn't really test much.

Templates provided for every programming assignment made this course quite a bit easier than it should have been.

Suggested improvements:
Discard the quizzes (or make them optional).
Get 1 or 2 "heavy-duty" programming assignments - no templates, you start from scratch.

Overall:
Good as a machine learning course, but great as an introductory course.

By Thomas Johnson from Chicago, Illinois 5 months ago Completed

One of the best classes I've taken. Ng provides excellent intuition for the algorithms he covers. The practical advice on designing machine learning pipelines in the latter part of the course is perhaps the best part. One downside is that some important machine learning techniques like decision trees and ensemble methods are not covered - I would love to see a "Machine Learning II" course by Ng

By Sergio Marchesini from Padua, Italy 6 months ago Completed

excellent course, great teaching, learning material was perfect. Andrew is an excellent teacher, plenty of code examples and real world applications...

By Robert Komartin from Bucharest, Bucuresti 6 months ago Completed

Excellent! I definitely recommend it!

By Ruslan Bes from Kharkov, Ukraine 7 months ago Completed

Prerequisites: some programming background would be good

Programming Exercises: maybe even too easy, almost everything is explained in detail.

Video Lectures: there are three main chapters — Linear/Logistic Regression, Neural Networks and several extensions and applications of these concepts in learning algorithms. The lectures usually have too extensive explanations so I watched most of the videos on speed 1.25.

What I've learned from the course:
- Algorithms of how to transform a lot of raw data into the meaningful statistics that allows to make a decision.
- Writing software that recognizes hand-written symbols isn't that hard. Same thing about recommender systems and spam-filtering.
- Side-effect: When one have to deal with arrays of data sometimes there is a better solution than writing a loop (vectorization).
- Side-effect: How to use Octave for simple mathematical tasks.

By Swizec Teller from Ljubljana, Ljubljana 8 months ago Completed

Took this course when it first came out and I really loved it. Prof. Ng is possibly the best person I have ever listened to explain a complex topic.

Was sitting a real life machine learning class at the same time and the two just cannot compare. The online class went through the material much quicker and focused more on things that are practical rather than things that were thought promising twenty years ago but have since fallen out of popular use.

By jledoux from nyc 2 days ago Taking Course Now

Great course if you lack math sophistication. Sadly though, for people with a good math background, he often skips any intuition for where things come from (e.g., with back propagation he just hands the equations and makes little to no attempt to justify how its derived). Or with "advanced" optimization methods, there's no brief lecture on what conjugate gradient method is, when its better than other methods, etc.

I also find it very patronizing when he says after 2-3 weeks of classes, he adds things like "You now know more about machine learning than most engineers doing ML in Silicon Valley" after teaching just simple linear and logistic regression.

(Prior to this course, I would have answered I knew next to nothing about ML, as I never consider curve fitting/optimization as ML. But from experimental physics courses knew how to use maximum likelihood and minimum of least squares, and the "advanced" stuff like conjugate gradient method of finding a minimum, that's far beyond this course. Granted I did learn a lot about Neural nets and SVM in a very accessible way).

However, it is a very easy very accessible introduction to the materials. I also find the programming assignments too spoon-fed. E.g., each part of the assignment leave out one line of octave code, and then give you the equation in the lecture which is trivial to switch into octave. Unlike other courses where you get a feel of "I managed to implement something on my own to solve a problem", you feel more like I can just change one relevant line in code written by someone else.

By Chris Beard from Atlanta, Georgia 23 days ago Completed

Can't say enough good things about this course and prof. Probably single most impactful course in deciding my career.

By Ilya from St. Petersburg 39 days ago Completed

Great course from Coursera's father. I can recommend this course to everyone who wanted to start learning ML discipline. Lecturer ( BTW one of the best scientists in ML area ) is very passionate about topic. He has a talent to explain complicated things in very gentle an easy manner.

This course is invaluable introduction to ML topic, and must be taken before more advanced courses like "Natural Language Processing" or "Neural Networks for Machine Learning".

By Jon Gauthier from Phoenix, Arizona 2 months ago Completed

-- What was your prior experience in the field?
None. (Web developer since 2006.)

-- What did you learn?
The Stanford ML course teaches a set of immediately applicable machine learning algorithms, from linear regression to feedforward neural networks. Professor Ng consistently includes in his lectures notes on the implementation of the content presented. He is straightforward about the caveats of the methods described in the course, and spends an entire section of the course enumerating the various ways to diagnose which errors are affecting a given implementation / application and how to make the proper correction.

-- Did the course meet expectations?
The course easily exceeded my expectations. The concepts in this course now serve as an entire new set of utilities on my toolbelt as a computer programmer. They have been enormously useful and have without a doubt added to my value as a programmer.

-- What didn't you like?
The most difficult math that was fully covered in the course dealt with matrix algebra. Concepts with steps involving calculus or linear algebra were only briefly described. While an understanding of the mathematical underpinnings is not required to build a competent implementation of one of the ML algorithms taught, it would have been interesting to see more (potentially optional) lectures on the more technical mathematical support that ML depends upon.

By xasmx 2 months ago Completed

An excellent introductory course to machine learning. It's heavily focused on practical issues of machine learning and after it you'll be able to use machine learning for your own purposes. The course does not cover theory of machine learning, so if your interest is more in the theory than in practice, you might feel that the course doesn't go deep enough. I prefer to first learn the practice and after that go deeper into the theoretical aspects, so this was the perfect introductory course for me.

By emediquei 3 months ago Completed

A great teacher and interesting content. Some of the programming assignments don't help very much in understanding the topics, and they just require filling in some blanks in Octave. But a good course nevertheless.

By Mayank Singh from Patna, Bihar 4 months ago Completed

A course that interested me in going for higher studies in this area. Couldn't believe I had created a handwriting recognizer. OCR was another awesome project. The projects weren't too tough but good enough for a beginner. The course is best for beginners wanting to experience the world of ML. I disagree with people about the course being watered down. It's meant for newbies and not an advanced course on the subject. Don't forget to check out kaggle when you're done with the course.

By Taqi 5 months ago Completed

The best course I have taken in my life ! This Dr is awesome ! The programming material is great and everyone can follow. Make sure to have enough time to complete the homework.

By Patrix Rembang 5 months ago Completed

This course is a good introduction to Machine Learning. You will be exposed to a handful of supervised and unsupervised learning algorithm. The professor really did a good job explaining concepts without assuming his audience have background in calculus or linear algebra. The programming assignments are fun, but not really difficult.

The downside of this course is the lack of math. If you're looking for hardcore or rigorous introduction to ML, you won't find it here. But if you just want to survey ML algorithms and some best practice advice, know some programming, and don't really know calculus and linear algebra, this is for you.

By Ethan Berl from Princeton, New Jersey 6 months ago Completed

This course is very good and well planned. Andrew Ng explains the material very well (albeit a bit slowly) and the content is extremely useful. The absolute most useful part of the course is that he focuses on how to tell when the algorithms are working and how to tell when something is going wrong. Over/under fitting, regularization, learning curves, precision vs. recall, etc. give a real insight into the subject rather than just handing the student a toolbox of algorithms which could be misused. The actual algorithms cover all the established techniques very well.
The one big complaint I had with this course was that the homeworks and quizzes were too easy. You were able to fill in the few lines of Octave code without really having to understand the algorithm completely, which to me is a fatal flaw and defeats the purpose of the homework. I was able to get full points on everything but I know that I would not be able to implement SVM in another language after the course -- even though I do have a reasonable overview understanding of what the algorithm achieves. Because of this hole, I can't give the course a perfect rating but other than this, the video lectures were excellent and the material is so useful I often refer back to it even though the course ended several months ago.

By Luka Kacil 6 months ago Completed

I really enjoyed Ng's course. He's one of the few professors who are also really awesome lecturers. And the course itself is very interesting.

By Alex Ott from Paderborn, North Rhine-Westphalia 6 months ago Completed

I took initial version of this course, when Coursera wasn't founded yet. I want to say, that I really liked this course - lectures & additional materials completely covered everything what I need to make it complete.
Andrew Ng has ability to explain complex things in very simple language, and although this course isn't so complex from mathematical point of view, it gave me enough background to start to dig deeper, into mathematical basics of ML and related stuff. Home works were very well designed.

By Chris Simmons from Vancouver 7 months ago Completed

This is the course that got me interested in MOOC's - it deserves 5 stars just for that, but it was a great course as well.

I took the initial offering of this course in Fall 2011, before Coursera existed. The material is interesting, covering a broad range of machine learning approaches. The programming assignments are reasonable if you have a computer science background, and would be much easier if you have experience with a data-based language like Octave, R, Matlab, etc.

One nit, and this is minor. Several times throughout the course, Andrew mentioned that learning the material presented in the course would put you above most ML users in Silicon Valley. Now that I'm in a company that does machine learning at a very large scale (albeit not located in the Valley), I find this assessment a bit questionable - these people really know their stuff.

Overall, a great course, and increasingly important in the era of big data.

By Marek Stój from Wrocław 7 months ago Completed

Very pragmatic approach to machine learning, the professor has great teaching skills. Programming assignments are extensive and very fun to complete.

By Ricardo Teixeira from Antwerp 7 months ago Completed

This was the best MOOC I took to date. Professor Ng has amazing teaching skills, particularly because he teaches such a hard class. The length of the lessons is just right, and the material he prepares for programming assignments is great because if guides you through the exercise.

Overall, a tremendous experience whether you have previous programming experience or not.

By Ben Haley from Chicago, Illinois 7 months ago Completed

Simply the best MOOC I have taken.

By Gui Ambros from Atlanta, Georgia 7 months ago Completed

Excellent introductory course for Machine Learning newbies. Covers a lot of ground in just a couple of months, from regression & classification, to K-Means, SVM, Neural Networks and more.

You'll be using GNU Octave (a free version of Matlab, but Matlab works as well if you have it).

I did the first class (the one in '11, before Coursera was invented) and was amazed by the quality and professionalism of Prof. Andrew Ng. Not a surprise to see Coursera now growing so fast.

We're definitely living the revolution of higher education.

By Ankur Sethi from Herndon, Virginia 7 months ago Completed (Partially)

This is a great class. The professor does give excellent videos and the material is very practical.

By Jesse Spaulding from San Francisco, California 8 months ago Completed (Partially)

I took this course in the Fall of 2011 and it's one of two courses that inspired me to create this site (CourseTalk).

Here's the best things about this course:
- Andrew Ng is awesome. He's a top expert in the field and you really feel like he's your personal tutor.
- The course makes machine learning very easy to understand.
- High production value.