Great course, although would have liked some practical assignment and hands on coding for the topics discussed.

more practice!

I'm glad I spent some time on the "Flight simulator" assignments in this course. It's the first time in the specialization when I actually found the quiz questions challenging, and that's a welcome change. However, I didn't learn too much from the lectures. They were too repetitive, either repeating themselves or the material from the previous course. One or two videos could also do with better editing work: I could hear Andrew making a soundcheck, and there's a 30sec segment that's played twice in a row. Overall, it's probably worth doing this course, given that it requires very little time, and the assignments are useful.

The 3rd course is more art than science. There is a lot of breadth, but we cover each topic in passing. Therefore, from a student perspective, I find that the concepts are not cemented and it is entirely possible that I forget them once I move on to the next course.

The second issue I find with the course is that there are no programming assignments. Programming assignments. Programming assignments are key to understanding such complex topics and getting the idea cemented. It would have been much better, if we could cover each topic such as data-mismatch, comparison to human level performance, etc via assignments.

While there is some new and interesting material, it is not enough to justify this being a course on its own. The material could have been easily added to previous courses, as it repeats or extends what was covered there.

The "flight simulator" quizzes can be challenging, but also boring. Some of the techniques presented should be practiced with programming exercises. Transfer learning or multitasking based on material from the previous courses would have been great. Knowing that these techniques exist is not enough.

A project teaching how to remove the last layer and retrain to learn a different task would have been very exciting and informative.

I completed the course for the sake of the specialization, but I was not very motivated.

this was definitely a useful course, as it addressed the 'art' of machine learning.

For me, the mathematics and writing code is easy - that's the science; however, it is equally important to have heuristics for deciding what sort of learning algorithm(s) to try, and how to start, and how to iterate.

That being said, some of the terminology is peculiar - satisficing, for example, is that even a real word?.

In the software requirements engineering field, we'd call that performance requirements ( for run-time speed), or perhaps non-functional requirements( memory usage), depending on the metric.

Also, in the second week, there was a discussion of error priorities for the autonomous vehicle example and quiz where a safety-critical requirement was not taken into consideration at all.

Spoiler Alert: If I am building the AI and control systems for a vehicle ( autonomous or otherwise), , that has to work in all weather conditions, no matter how hard it might be to get the necessary training data. Qualifying the answer with 'all other things being equal' never applies to safety-critical systems.

First of all, I thank Professor Andrew Ng for offering this high quality "Deep Learning" specialization. This specialization helped me overall to gain a solid fundamentals and strong intuition about building blocks of Neural Networks. I'm looking forward to have a next level course on top of this track. Thanks again, Sir!

I strongly recommend this specialization for anyone who wish get their hands dirty and wants to understand what really happens under the hood of Neural networks with some curiosity.

Some of the key factors that differentiate this specialization from other specialization course:

1. Concepts are laid from ground up (i.e you to got to build models using basic numpy/pandas/python and then all the way up using tensorflow and keras etc)

2. Programming Assignments at end of each week on every course.

3. Reference to influential research papers on each topics and guidance provided to study those articles.

4. Motivation talks from few great leaders and scientist from Deep Learning field/community.

I really liked to learn about the actual problems faced in a project and the ways to tackle them more or less systematically. I also understood the challenges and open questions in case of dead ends. The two quizzes really can help me answer a typical deep learning job interview. I definitely feel prepared for a job in deep learning industry. Finally, the interviews with Andrej ( I have read his blogs but never got to see a video/picture) and Russ were thrilling and keeps me motivating to not approach deep learning as a subject solved but an evolving research area. It also tells me to revisit some of the concepts like autoencoders, RBMs which are normally not dealt in normal deep learning class. Once again, I want to thank Prof Andrew for his simple, elegant and thought provoking lectures which are not only specific but also fulfilling. It is extremely interesting to do his course just like watching a favorite movie/ series. Thank you Coursera team !

I find this content really interesting and useful to gain experience you need in this field.

I couldn't believe I failed two times in row for an assignment and finally passed the third time with a grade of 80%. :O (I also tried my best to not memorize any question even though questions were gets updates on each try!) Part of my failure was caused by the fact this martial is easy to learn but hard to master as prof. Andrew said. The other part is because you could have different interpretations on the questions and they could get ambiguous with a little more thinking.

I say they did really great on making this kind of assignment. Genius idea!! though I wished there were more clarification on questions and more careful about choosing the words. e.g. the sentence "Having 3 evolution metrics" could be different from "Using 3 evaluation metrics".

Thanks professor. Anderw and to all contributers of making the Deep Learning Specialization.

This is by far the most useful and unique course on Machine Learning you will find around. After many years as a research engineer, I have not yet seen anywhere a better set of systematic approaches and guidelines about how to go on with your ML / Deep Learning projects. Here you are not gonna see almost no math, but the videos are packed with condensed knowledge and years of experience from Andrew Ng. The part of Error Analysis shows a deep understanding and knowledge of him about the intricacies of DL development. This is a course on how to do Deep Learning, not much about the models and the data, but how yourself to use the results of the experiments in order to progress in building better ML/DL systems. It is almost a type of optimization procedure you are gonna use not to train your model, but to train your team and yourself on how to achieve the best results in your Deep Learning and Machine Learning projects.

I'm Zeyad, an undergraduate of Computer Engineering at Alexandria University in Egypt.

Taking this course really helped me to learn and study this field and also to implement it. It helped me advance in my knowledge. This course helped me defining Deep Learning field, understanding how Deep Learning could potentially impact our business and industry to write a thought leadership piece regarding use cases and industry potential of Machine Learning.

This specialization helped me identifying which aspects of Deep Learning field seem most important and relevant to us, apparently they were all important to us. Walking away with a strong foundation in where Deep Learning is going, what it does, and how to prepare for it.

Deep Learning specialization helped me achieving a good learning and knowledge about that field.

Thank you so much for offering such wonderful piece of art.

Best Regards,

Zeyad

Amazing courses that go into detailed explanations about the math and intuitions behind the algorithms without getting too convoluted or making things unnecessarily complicated just for the sake of it.

Prof. Andrew doesn’t just tell you the name of a function for a library (like scikit

learn or tensorflow) and give you magic numbers for parameters. You actually build the model yourself and learn what the parameters stand for and what is the purpose of those parameters and hyper-parameters.

The specialization is well divided into meaningful courses and each course is well structured so that you know exactly what you are going to learn and what key specific skills you will get after completion of a course. Because of the quizzes and practical labs, after completing a course you actually gain confidence that you can design optimized solutions for that particular set of problems.

Failure on a beginner level quizzes it's very irritating, more specially to me. I don't regret seen myself have to re-do some quizzes for every week, probably that's because English is not my first language, or how can I put it...my mother's tongue. I believe it's an indication of our weaknesses and if we face them we can grow to prosper, not that I am trying to be a life philosopher. Questions on this course are made in such a way to test if you really understood what the instructor has taught you. I love Andrew's ways of teaching, I just wish he was my electronics lecturer. I feel like I could have understood some of fuzzy concepts that I battled with very easily. The concept were given in such a structured way and I was very excited in many of these teaching and insights regarding machine learning approaches as a machine learning engineer.

This course is radically different from the first two of the specialization. While before we were dealing with the theoretical basis for how learning works and ways to optimize the performance of the computer, this one is more like a stream of tips, cautionary tales, and hacks in order to optimize the performance of the human. Personally, I found the material to be very educational and engaging, with many "aha" moments when the instructor makes you see the "obvious" solution for a problem that just seconds ago seemed unsolvable.

The assignments (the "flight simulator") are incredibly useful and make you think profoundly and systematically on the problems. I found that the questions would typically prompt even more questions in my head and make me consider many options to tackle a particular problem.

After seven days learning, I finally finished the three course of this specilization. I've gotten much more than I've expected at the beginning. Not only deeply understand how the neural network works, but also how to build deep neural network and how to train it efficiently. Now I know how to start to build a machine learning project and solve the specific problems from data preparation to model training and I know how to quickly get my network works through transfer learning and fine-tuning, etc. By watching the interview videos I got a lot about the future of AI and I deeply know what I am really interested in now. I really appreciate what Prof. Andrew and TAs have done to make this series available from all around the world and I really too impatient to wait to learn the next two course.

This is probably the most important course in the specialization. It's very easy now-a-days to create Neural Networks and get a grasp of how they work due to high-level frameworks (keras, scikit, tflearn, etc) and abundance of literature and videos, respectively. The thing that is lacking from most resources that I have encountered on learning Deep Learning and Neural Nets is how to optimize and approach problems. I have in the past build some complex Neural Networks, but would hit road blocks that would ruin productivity for I didn't know how to approach problems correctly, and didn't know what knobs to turn to improve performance of my program. This course teaches techniques that I find extremely useful for my previous problems in Machine Learning.

Very useful knowledge, super interesting material and prof. Ng is an awesome teacher as always. The simulating approach for the quiz is great! However the "simulation" questions and answers should be carefully reviewed. Sometimes the "right" answer is difficult to choose because of an ambiguity or a little detail that does not quite match the lectures and two answers seem to have some of the right element OR no answer seems to be perfectly right. Going thru the forums, you will find plenty of comments like this to figure out which questions to tune. Some are right and some are due to the student genuinely making a mistake. Perhaps looking at the error rate on each question will also help seeing which one are abnormally incorrectly answered.

Loved the course because the insights shared by Andrew Ng are clearly coming from real-world industry experience. Besides the content of the video lectures, which are a must-see for every ML practitioner, I particularly liked the "flight simulator"-style assignments.

Although the content is of very high quality, I noted that there a couple of mistakes in the assignment texts, unfortunately sometimes even in the options of multiple-choice questions, which make it unnecessarily hard to guess what the option actually means. In one case (assignment 2, question 10) I even think the "correct" answer's text is contradictory to what Andrew says in the lecture. I feel that half an hour of proof-reading could have taken care of these mistakes.

Following this course resulted in a tremendous boost in my machine learning experience, especially that I was just starting a new ML project and I was able to practice every week the techniques suggested by Andrew. I recommend to everyone that you start a fresh ML project and apply these guidelines, even if it is just a toy project or Kaggle competition for example. One additional encouragement is that these advice helps a lot in the context of structured data as well. I am working on optimizing computer network traffic and by considering the information Andrew presents I was able to influence not only the ML progress on the project but also make the rest of the teams reconsider their pipeline. Love these simulators.

Previously, I have taken online classes before in Machine Learning by going the cheap route (Udemy, blogs, youtube) and you get what you pay for. Andrew Ng explains it the most thorough, easiest, and simplest way possible. Presentation material is very understandable. Great class for new machine learning learners. Highly recommend it. The only downside is that the programming exercises are little too easy in my opinion. I feel like the best way to get your hands dirty is to do actual projects (do your own projects). These lectures are good for intuition and background of different types of Neural Network architectures. Other than that, Great material. Thanks Andrew!

In this course, I learned a lot about how to make right decisions when facing different problems in machine learning tasks. It helps me to review the decisions I made in the past, and also shows me a more systematic way to think about what to do next. I strongly recommend everyone interested in ML to take this course.

The only thing I'm not so satisfied with is that some questions in the quiz are quite confusing. Maybe they just have wording issue, but these questions and their corresponding answers do confuse a lot of people. I think maybe TA could take some time to address these problems in the discussion forum and help us learn even better.

Another excellent course made by Andrew Ng. It is another perfect example of how to prepare good learning materials.

This course does not in fact expect you to write the code. Teacher is aiming not to offer you his abilities to make working system. He is offering you his deep insight and experience in making systems better and better to the point in which they meet expectactions. He discusses how to address issues you may encounter in systematic manner and where put your resources to use them in most efficient way.

If you are building machine learning models I am sure that this course pays off and can spare you many mistakes you could make.

This course is perhaps the most under-rated one among all the courses under this specialization. After working in the field for a while now, the most common bottleneck my team and myself have encountered is often related to inadequate understanding of data sourcing, data distribution, training results and error analysis.

Course 4 and 5 are good basic introduction to computer vision and natural language processing since they cover some of the most important topics and Andrew Ng provides rather useful "intuitions" around them, but there's good reason to make this course one of its own (despite that it's only two weeks long).

This is a passive course. Don't let the 2-week course set you off. The videos in here are really insightful. They give you some of the experience that Andrew has seen throughout the years.

They will provide you with the right way on how to split the data sets, how to handle when the train, dev & test sets come from different distributions; advantages of orthogonalization; The avoidable bias, the satisfying and optimizing metrics.

By investing in this course, this will save you tons and tons of hours of work by understanding some key concepts that you will need for an effective Machine Learning problem.

An amazing course indeed. A bit "dull" to some due to the lack of programming assignments, but extremely beneficial and insightful to anyone seriously considering to tackle an ML project. You have to appreciate the fact that while what this course covers may sometimes seem like "common sense", it is still reassuring and comforting to know that these concepts and principles are what the likes of Prof. Andrew Ng go by when they embark on an ML project.

To all who are working on making this platform what it is, I'm very confident that it is not an easy thing at all, so thank you so much.