Nice

good

9

It's an introductory course, so what you'll get is an intruductiory overview. During the lecture videos, you'll have to take some things for granted. Some of them are explained later, some are not. What I do mind is that there is no interaction between the course staff (lectures or assistants) and course participants. So some of your questions will be left unanswered, and on some you'll get questionable answers. More and more I find this to be the general problem with Coursera. You have few graded quizes and few lab exercises. So in my opinion, the course is not worth paying extra money for the certificate.

I enjoyed Lawrence Carin's explanations a lot but the overall experience was not great I'm afraid. It felt like it did not come together properly. The order of lectures and assignments felt wrong. The Python level of competence was too high for this course and support (via forums) was non-existent. I don't regret finishing the course but I would not recommend it to my friends.

The lectures were great and very easy to follow! However, I found that the assessments were too easy as they comprised solely of multiple choice questions, maybe including hands on coding assessments fo contribute to our final grade would have made me feel more confident that I've grasped what was supposed to be taught to me each week.

Modules need a bit of synchronization.

Please spend some more time explaining gradient descent.

If possible, explain using a board where we could have things simultaneously.

Also, request to have a course on machine vision using CNN etc.

Good introduction to the concepts of machine learning. Somewhat overly repetitive, some inconsistencies. Information is not up to date, for example, the python code for Tensorflow would not work.

This course is good for just theoretical understanding of the subject. But for practical implementation it is too hard to do.

Thank You team Coursera and Teachers from Duke University for helping me to understan dthe basics of machine learning..

More examples should be provided for each course, and the content needs to be more simplified.

A bit too technical for new learners. More practical exercises will make it more learnable.

The theory is well explained but you guys should update the coding parts to TensorFlow 2.

There are many better ML intro courses out there...poorly structued and delivered.

Only concepts no practice. But the concepts are really well taught according to me

The course requires more detailed intuitive approach towards material preparation

some concepts were difficult to understand and not explained very well

just a basic overview of the methods. not much worth the time

I could not even understand the introduction class

Pretty good introduction to Machine Learning!

4.2

Although the teachings of machine learning are introduction. A learner needs to know moderate level of python and pytorch to continue the course fluently.

The concept is taught ok, but it's not great on hands-on learning.

Topics are explained in a random and ilogical fashion. For example, they teach CNNs before gradient descent or the basics for training a model (like training splits or criteria to evaluate a model). Lack of order, toghether with confusing figures and poor metaphors, make impossible to properly understand any concept, and that can be seen in the discussion forums. In addition, there are gross conceptual errors, like suggesting that problems with non-linear solutions requiere NNs and can't be solved with linear regression algorithms. In fact, the whole course is centered in neural networks despite being presented as an introduction to ML and sells the idea that somehow NNs are the ideal solution to any non-trivial problem. This course is in shocking contrast with some other excelent courses that can be found in this pplatform, like the ones from Standford or Michigan.