b

.

g

The course aims to describe the key elements and techniques of machine learning but without delving into the associated maths. It manages to achieve this. The first lectures are delivered in a somewhat hesitant way but they improve over time and provide some very good insights as to how the procedures like convoluted neural networks work. Having only previously had exposure to Andrew Ng's Introduction to Machine Learning Course, I thought this course was a useful complement to the Stanford IMLC course.

So why not five stars? There were a few annoying things that could be improved IMHO:

1. The course is six weeks long, but there are only tests for the first four weeks. Once you complete these, you get messages saying you've completed the course, despite there being two (rather important) weeks left...why not have tests for the last two weeks of content??

2. The computing assignments are not graded, but there is no prior indication that this is the case! Probably just as well, since they assume a degree of familiarity with Python and Pytorch that is beyond the casual computer-literate person (though the Pytorch website provides lots of relevant help in this respect!) but it would have saved lots of angst to know that these were in effect optional.

3. The discussion boards are not reviewed by the course providers; numerous student questions are left unanswered, unless other students help. This is in sharp contrast to the Stanford course, where the discussion boards are very active and tutors are on hand to help and provide guidance.

But all in all, a useful and enjoyable course for a novice like me.

Overall a good course. The explanations are thorough and understandable, modern methods and applications are discussed, and some valuable code is provided. That being said, there are several things I feel are important to point out:

- This course focuses heavily on neural networks, and application-wise focuses heavily on image analysis and natural language processing. So it's not a true introduction to general "machine learning," but rather to neural networks. If you're not interested in those methods or the aforementioned applications, I'd find a different course.

- There are coding examples provided with simple written explanations, but otherwise this course doesn't focus on coding so much as discussions of the actual methods and algorithms. If you want to learn deep/machine learning coding libraries, I'd find a different course.

- Based on the discussion boards, it doesn't appear that this course is actively maintained as of January 2023. The videos and concepts are still applicable, and the coding examples still work, but there's no active feedback or communication from any Duke staff.

Like I said, still a good course (great if you want to learn neural networks). But the content covers very specific topics.

I am a beginner in ML/Neural networks and am a 1st year college student and know only basic python and a little more than high school math. This course was excellent for me to understand the basics of so many fascinating topics like CNN, RNN, NLP etc and is helpful for coming up with ideas for my mini project.For anyone taking this course, make sure you have a little bit of background knowledge of what is to be expected. Just some familiarity with how neural networks resemble the human mind , how solving things using a human and predicting things using a machine is different etc.

However, I wasn't able to complete any of the lab assignments. Not sure if it is because of my lack of knowledge of python or the method in which the lab portion was dealt with in the course. My opinion is that videos explaining the lab work would have been more effective that just having to read the documentation.

It's a really great course for learning basic concepts of machine learning/deep learning. Professor Carvin Lawrence put a great effort in making this course fruitful. Although there were few things which can be improved like the coding part. In coding part we only have Jupyter notebooks and no video lectures at all. It would be great to replace text files of code to proper video oriented programming lectures. I think if the above mentioned problem is resolved then the course will definitely receive 5 stars from everyone.

I recommend this course for the people who are interested in Machine Learning and have no knowledge or beginner knowledge. The concepts of ML are explained in very clearly and in detailed manner. If someone is looking only for the coding part of Machine Learning, sorry but this course is not for you according to me but you can also check out yourself.

From a theoretical standpoint, the course fulfills its purpose giving you a nice introduction into Machine Learning as whole. It pretty much goes through all the famous models there are available nowadays, so you can ride comfortably in this word (at least you'll know what an RNN/LSTM, etc. are when you hear it).

Unfortunately, from a practical standpoint, it is very lackluster. Theory and practice just don't link. You're given lessons on Jupyter notebook that are just disjointed from the explanations, sometimes requiring resources you're not even introduced to yet (and are actually talked about AFTER you do the lessons, like in week 6). You are definitely going to need some extra online resources to fully understand how to actually implement some neural networks by yourself.

Overall the course is great, materials are fundamental to anyone pursuing a journey into the world of AI through Deep Learning, the course should be named introduction to Deep Learning, rather than Machine Learning which often involves treating a wide variety of more classical algorithms not discussed in the course. However, the course has its own advantages, being a hands-on approach, short and brief right to the point, a fairly great introduction to the little history of Neural Networks which really puts us into perspective, and great instructors. The quality is fine. Still lacks more hands-on experiment especially on leveraging the true power of Transformers. Hence one star lost!

The lectures were better than outstanding, they were masterful. A logical progression, easy to understand, setting the student up for success. It's amazing how 50 years of research in the field was concisely summarized into six weeks at an introductory level.

A logical transition to the labs and introduction to the libraries used, however, was severely lacking to nonexistent. I spent more time on the Lesson 2 lab exercise than all of the lectures combined and was not able to write a single line of code. An introduction to the labyrinth of the numpy library could have prevented that exercise in frustration.

It is a good way to get to know what the latest state of the art machine algorithms are like, though I wouldn't recommend this course to anyone who is just getting into machine learning. It can be confusing at times because of notation and typos, still the course content is interesting and up to date . Regarding, implementation using pytorch is great because it allows to develop complex model, but is kinda complex and sometimes might seem like a black box even for python savvy learners; that's why I think there should be more emphasis on explaining this in more detail.

Four stars as it was offered to me (with certificate) for free. Had I had to pay for this it would have been only 2* Despite being advertised as an Introduction (and much repetition in the videos as to how basic the content was supposed to be) I felt the course was considerably above my level. A strong computer background and good coding skills were required as a prerequsit to get much from the course.

Good for those already at a higher level, but beware this is not an 'Introduction' for those who are completely new to the subject.

The content is what it says. Touched on both practical and theoretical aspects. The only problem: for people who want to do theory it does not provide a solid math foundation, and for people who want to use it their examples are way too brief, one of them doesn't even work (the code for text processing part are outdated and doesn't run even on their lab).

I think it does have good content although they really need to polish a bit for a 5-star.

I really enjoyed this course. I like how the various instructors have explained the concepts with simple words and using numerous examples. I would not say that this is an introductory course but more a middle one since it focuses more on deep learning. I would suggest this course to anyone who has already done some machine learning and want to revise some basic concept of ML and have and introduction on DL.

It is all in all an interesting course but I am glad I did a more basic introduction before. It is not really for beginners and it is , at times, hard to follow. The worst part are the exercises. They are there, but you are pretty much on your own and as far as I can see there are no solutions or hints. The instructors appear to know the subject but the explanations are not always coincise.

Explanation about the Model Infrastructure is too good, an eyeopener and highly informative.

Lab work assignments can be made better if their difficulty level should be increased step by step and their should be some hints how to find the answer for the given questions or correct answer/methods should be provided after lab notebook submission so that student can understand their faults.

Very good content just that for me it is outside what I expected. Looks to me this is more on NLP than using algorithms like KNN, Random Forest and others to make predictions and talking about Supervised and Unsupervised Learning

Personally, it is good introduction to what I have being dreaming about

Some topics are covered well at an introductory level. Some material was repeated several times for memorization. Some topics are not covered very deeply. In some of the topics, there was little detail. There are few formulas at the beginning of the course. Some topics explain not very deep.

It was a great experience learning with the professors of duke university .As far as this course is concerned the concepts of machine learning have clearly explained by the the mentors which will be helpful in my future. Thanks Coursera for providing me such good opportunity .

Teaching methods are good, explanation is also detailed and very informative but if possible to give practical session and teaching about how to do Hands-on, how to code the for machine learning, it would have been more better. Hoping for consideration of this small request...;)

Really good course, I've learned a lot of new things. The course changed completely my initial ideas about Machine Learning Concepts and Utility. Also, it was really useful to situate ML relation with Neural Networks, Deep learning, IA, etc... Totally I recommend it

Good course and practical samples with good depth and looking beyond the implementation of tools, especially regarding the structure of algorithms and complex mathematical concepts behind the functions and parameters in Machine learning models. Thanks professors!

this course was very organized and well teaching' from basic through the mathematic principles and guidelines till the logical implmentation and the complicated architectures for getting these algorythms into practical.

thx a lot for the inspiring course,

Orel

I like the course and all the video materials. However, Week 6 Lab seems to be out of place. It is dependent on topics that are introduced after it again. Sometimes, topics get repeated several times as if it is the first time they are shown to the student.

This is a short and quick course to get one kick started with Deep Learning. It covers basic building blocks for 2 very important tasks - Image Analysis and Text Analysis. All instructors have good command of the subject and teaching method is superb.