I have given this course four weeks and am officially firing the instructor. I find the videos unintuitive. They lack a description of concepts and rather focus on the mechanics first when I don't even understand why this matters or have a conceptual framework with which to undersatnd it. I have taken to Youtube to learn the concepts, remove the consuion caused by the Coursera videos and come back to take the assessment on Coursera. I would suggest deeplearning.ai hire 3BlueOneBrown from YT to develop proper videos. The workbooks often test me on things I expect the course to teach me first. For example, there is a graded assignment on neural networks in which you code a neural net. Yet this info is not provided in the videos, and the description in the notebook is meager at best. I have spent over 30 HOURS on the workbooks trying to decode what they are asking me to do. I manage to finish each but It's no way to learn. If you plan to take this course, I would strongly suggest clicking ahead and sample the videos and labs in detail and decide if this is for you. Oh, and put Python as a requisite on the description page!

It breaks my heart to only give this course two stars, but some things need improvement. The videos are masterfully explained and I loved every second, but the python assignments are very buggy, especially the one from week 3.

I believe that the main problem with these assignments is that they are a bunch of code that you can't understand absolutely anything. I strongly recommend that deeplearning update this codes with comments on each line so that we can understand step by step what is being done. Also, I went through several situations where the value was correct, the cell test was green, but the grader gave me 0/100.

The errors menssage from the grader do not explain the reason for the error at all. Finally, the deeplearning forum is dead and the administrators either ignore or just don't bother to answering the posts.

Pros:

1. Concepts are explained in simple terms and complexity builds slowly over time. I felt more confident after each video rather than confused like with other courses.

2. Each video has small quizzes to solidify your understanding of small topics on the spot

3. Solving weekly quiz questions with pen and paper helps build mechanical memory.

Things that could be improved:

1. It would be nice to have a pdf file with all formulas in one place to refer to.

2. I wish there were a bit more examples of eigenvalues and eigenvectors, I had to do external research to fully understand the topic.

Overall, great course for beginners and those who have already started learning ML and want to get better intuition of math behind it.

Too basic and chaotic

They teach banana and they ask to solve assignment with python

Do not waste your time

They already wasted my time for 2 weeks

So bad course

Overall, I think this was a good introduction to linear algebra. It was fairly easy to follow along and complete assignments. I learned a bunch and got some new perspective on my overall journey as data scientist. If you want to be successful with this, I would recommend actually doing some extra practice manipulating matrices, playing with the different dynamic graphs this course offers to get some more intuition with how things move and possibly doing some supplemental learning watching youtube videos on anything that gave you trouble or checking out the same material on Khan Academy (recommendations they give as well).

Things became really confusing to follow around the eigenvalues and eigenvectors. I had to rewatch the videos a handful of times to try to piece together what was happening and how. There was a couple "Tada!" moments and I wasn't sure how we'd gotten where we were. When it came time to do the quiz, I had a couple questions I had to guess since I couldn't make the connection from what I was lectured on vs. how to complete problems I was seeing.

To be fair, I could have leveraged the forum more to ask questions and gain clarity but you have to go to a different website, register another account, scroll past the first page of announcements and general discussion, reverse the order of the thing you're looking for to find the math specialization header, click into that to find the M4ML Course 1 (not the Linear Algebra course that my brain is looking for) and then the specific week you're in. As a Coursera user, there are some basic things in certain places that I expect to find them. I'm sure there are reasons for everything but this course lacks a key integration for a classroom when you have to go on a small journey that requires more personal information be doled out just to have a conversation with your peers and teachers.

The section on Eigenvectors is very fragmented and clueless. concepts are randomly arranged and not connected.

please explain the basics better

everything was very great and wonderful for the material except for the eigenvalues and eigenvectors, it feels so off and lacking in details, luckily i'm an engineering student who had studied about linear algebra before so i was able to follow trough. well if you are a new comer for this field, i think you should prepare more for the eigenvalues and eigenvector materials

I enjoyed this course very much. Every concept was well introduced and explained with examples. It's the best for beginners who want to algebra for Machine Learning.

Too simple

This course is good for you if you are beginner in mathematics knowledge in Machine Learning. Special feature that I like from this course is you will learn how to integrate mathematics formula into python programming language easily. I am not really good in python language but their instruction in their programming assignment is easy to understand and guide me step by step that I can finish all of their assignment. Furthermore, the community feature is also helpful so that I can finish my programming assignment as well.

Overall, when I take this course, I don't only learn mathematics in Machine Learning but also python programming which is beneficial and unique experience for me.

This is too light.

I did learn from this course but I feel there's a big room for improvement.

Finding the community/discussion page for this course was confusing and it took more clicks than it should.

Quizzes were introduced a lot in the beginning but I felt they were lacking in the middle weeks.

My most important feedback is that I feel the course did not adequately explain the applications of the lessons we were learning. We need deeper explanations/comparisons/illustrations showing how the mathematical concepts are applied in ML, I know there are labs but I felt they weren't enough.

Instructors aren't replying to many of the questions posted in the forums which their answers do in fact matter.

The answers to the quiz questions are sometimes not explained, you do have it 90% of the time but sometimes it just shows "Correct!" which does not show the work done for the answer. , also linking each question to the lectures would much appreciated (I know sometimes its there but not always.).

Something I would love to see is how about some math exercises for each concept explained? That would greatly cement the ideas taught to us.

A lot of the math formulas are brushed over and almost as if the instructor assumes you already know how to solve the problems. There isn't a lot of content on the "how". How does the instructor calculate transformations, eigenvalues, etc. None of this is taught - it's just shown more like a result.

Two thumbs up! Linear Algebra for Machine Learning and Data Science by DeepLearning.AI definitely makes for an excellent beginner's course for appreciating the significance of Linear Algebra in ML and DS. Kudos!

Best math refresher course. I like this style of the course structure. Starting from basic and intuitive examples, but following the steps I realized I understand all-important concept

Very helpful and initiative course for who has a good mathematics background or not.

Excellent course, It starts from zero and explains really complex concepts!

User beware - course programming exercises are structured quite poorly. First graded exercise (Gaussian Elimination) is puzzling in its design where it is (1) difficult to follow the logic of the function (different ways of doing reduced row echelon form) and (2) very difficult to trouble shoot what is going wrong given it is set up as a compound function. Not a newbie to programming or matrix math, so I wouldn't chalk it down to this. It's a shame as the course curriculum and structure is otherwise quite interesting.

My recommendation for those who are considering paying for this specialization is to search up 3Blue1Brown and learn from him FOR FREE. The concepts themselves are not bad albeit being a little shallow and surface level. The python labs are a joke because the auto grader does not work and no one in the community will help you resolve it.

After almost 2 weeks, I found that linkages to ML are weak and unclear (of course, there may be some. but I expect the instructor should demonstrate more) and Python programming projects are (and will be) coming out of the blue so I can't continue. Very disappointed.

Could have just mentioned it as linear algebra. Very few cases where ML is involved in the lectures. A few good youtube videos could have done the same job in even less time. Waste of money.

Quite useless course. Teachess really really basic concepts. There are no details. All excersies are designed as a simple plugging in of numbers.

It is the most interesting algebra course I ever took. The apple and banana system is really clear and simple to understand vector transform.