임페리얼 칼리지 런던의 Mathematics for Machine Learning: Linear Algebra 학습자 리뷰 및 피드백

This course was like God-gifted.

I had just finished my 2nd sem at college(BTech) and we had Matrices in the syllabus so I knew how to calculate (just calculate -_-) eigenvalues, vectors and so on but I just saw them as numbers. At my college, we were not given such geometric insight and when I learned it through this course, MY GOD was I blown away.

I feel so lucky to have found this course! I learned A TON of stuff.

Thanks!!

I took a great pleasure to study this linear algebra course, teachers are very talented since their way to explain mathematical concepts make it very easy to understand , in fact with this particular amazing approach I changed my perception about learning math and sciences in general. I do recommend this course if you look for a global overview of linear algebra for direct application in machine learning or computer sciences!

This course is great for those who want to understand the geometric meaning of linear algebra. Really loved the course videos and quizzes. Just one suggestion - Coding assignments should be bit more challenging as this course is targeted around ML, maybe some small Kaggle like project at the end of course.

Excellent, crisp and to the point. Instructors made the concepts way to easy to understand. Enjoyed my time learning from them and ofcourse relevant material was provided.

A good value, well organized, with many exercises for practice. Effectively uses visuals, and contains the occasional very creative example.

Some caveats

a) this course is not for the absolute beginner. You'll need secondary / high school math, and basic familiarity with python

b) understanding linear algebra at this level is a second year full semester course at university. So if you want to understand the concepts - rather than just get the certificate - be prepared to use outside resources and invest considerably more time than advertised. Some linear algebra topics are skipped (cross product), and others are not well integrated into the course (Einstein summation)

c) while linear algebra is central to understanding machine learning, there are very few machine learning applications in this course.

And finally a small annoyance: I wish the instructors would get out of the way of the whiteboard at the end, so I could get a screen capture.

Overall, a worthwhile course.

D

totally enjoyed it. requires a bit of side research as any online course would. some of the quizzes were not directly related to the video that preceded them as one would expect. However, a fun course and covers a lot of important basics for it's relatively short duration.

This course reviews the essential concept of linear algebra in the context of machine learning. However, it would be much better if it provided more optional exercise and reading materials.

decent course. It gives a good enough background to understand the mathematics necessities of many areas of data science. could be more thorough and dive deeper into some of the content.

Good course because it shows how to understand geometrically, things that I had hitherto only understood computationally.

Excellent Instruction

For the most part, I enjoyed this course. Most of the math explained is fairly easy to understand. They cover the fundamentals of linear algebra, and provide plenty of assignments and practice exercises to test your knowledge. However, some of the video explanations are extremely confusing and feel rushed. For example, some videos in Week 4 and 5 like Reflecting in a plane using Gram-Schmidt and the PageRank algorithm were so hard to understand that I had to learn about them from elsewhere on the internet (I used MIT OCW a lot). This isn't very convenient, especially if you're paying for the course. Furthermore, I felt like more videos explaining the applications of linear algebra in machine learning could've been made, and the ones that were already made could've been made in more detail (for example, the term 'span' was never formally explained). Lastly, I would've loved it if there was another week dedicated solely to introduce the coding bit, because it's really difficult and takes a while if you have little or no prior experience in python. All in all though, I enjoyed this course, and I would recommend trying to complete both Linear Algebra and Multivariate Calculus in one month, because it's not worth paying more than that.

Challenging course. Much more difficult that I expected. It took me 7-9 hours a week. The overall course material itself was good building-blocks to further understand application of machine learning. However, explanation in some topics should have more detailed explanation and examples to further understand the concept. There were many times, I need to re-watch each video over and over again, paused it, and figured things out on my own. The programming assignments were the most challenging task. I just began to learn Python and found it very difficult because there were so many codes I haven't learnt before. I think for those who has not learnt Python at all may find really really difficult to pass the assignments.

The content was good, but a couple of times what was said didn't gel with what was being drawn/written/done. Since I'm learning, this took me longer to double check when I misunderstood something whether it was the concept or a mistake in the delivery.

The topics could be improved in the way they are presented. I always had to search for additional material.

However, the course is okay, it could be better, the forum is not that active, and some assignments are good.

many of the activities are excellent, but videos hard to follow along to at times - play them at 0.75 speed if you can. Also, the faculty is not super responsive it seems on discussion boards creating some confusion

Although the course content is good, I feel it could have been done better. I enjoyed the multivariate calculus course compared to this course.

Some exercises are completely incoherent to the preceding videos, which makes it very difficult to solve them. very frustrating

Assignment makes frustration and not able to concentrate on teaching content

The course tries to cover every edge of Linear Algebra but fails to integrate each step with what relationship it has with Machine Learning. Core Formulas and Mathematical derivations are shoved down from throat without any respect for learners from non-engineering or computer science background. Other than week 1,2 rest was completely case study or example less UN-intuitive lectures of matrix formations and transformations. Needs a severe revamp with better examples and broader picture.

Mainly explains how to operate with matrices and vectors. Not how to use those in machine learning. If you expect to have a clear view of the usefulness of eigenvectors and eigenvalues in machine learning, this is not your course.

Some concepts were not clearly explained and there were a lot of issues with assignment grading working properly.

Not good, concepts not explained clearly.

Does not explain mathematics in videos

Authors try to teach babies. Might be good, it is hard to judge for me as I know linear algebra. Definitely boring to me. For example 3Blue1Brown (which they reference btw) is ingenious in my opinion, so it might be not me who is the problem.

But the quizzes just don't make sense! The ones where solving problems involved might have 2 numerically right answers but only one of two is treated as the right. And there are just idiotic or not covered in lectures answers for quizzes without problems.

Hands down the worst course I've ever taken in my life! Poorly put together and extremely short videos that don't provide an adequate amount of knowledge especially in relationship to the given quizzes. I truly hope this course is removed.