very good course

I have just finished the old machine learning course, and I'm doing this because I'm learning python/numpy/matplotlib. I thought the question during the course and quizzes insulted my intelligence. The material is great, but you need to improve the simple questions and quizzes. The first programming assignment was too easy, the second programming assignment was at a fair level. I still think more should be left to the student to do.

The quizes are too straight forward and simple. The code exercise too short as well.

Also disappointed that vectorisation is introduced but cost and loss functions are still calculated in for loops.

Brilliantly Designed course to teach beginer on Machine Learning. The course focuses on the theory behind machine learning. The content convered in the course allows the student to get an intuitive idea behind machine learning and gives him an idea of the mathematics behind it. The course is not very math intensive, but there is just enough math covered here to give the student an intuitive idea of machine learning.

The coding labs provide very detailed code, which the user can learn and analyze to make his own machine learning algorithm

My favorite part about this course was how neatly the jupyter notebooks and python files of the lab were arranged and provided. These lab files take the burden of coding from scratch away from the students, and allow students to focus only on the algorithms behind machine learning.

After this course, machine learning codes will no longer be a black box, but will be something you will understand very well. So, after doing this course, the next time you use Machine learning libraries like SciKitLearn, you will know exactly what is going on behind the curtains, can you can adjust parameters of ready-built ML funcitons to fit your needs.

At the end of this course, you will learn how you can modify machine learning codes for each custom need, and you will gain the ability to do those modifications yourself. After this course, you will be able to write specific machine learning codes which are well suited for a different specific application

The course is very easy to follow, building slowly enough and with enough examples that it's usually simple to understand, and then, looking back, you discover that you have learned something quite complicated. I have enough basic coding experience in python to handle basic functions such as those in this course already, so I found that part quite easy; this may not be the case for those with no python background at all.

Andrew Ng has a great teaching persona, and it's a real pleasure to watch the videos, even aside from what I'm learning, just because the vibe is so cheerful and supportive. As an educator and teacher trainer, I can be quite critical of how courses are taught, but this one is just a joy. I feel like there's a lot for me to learn from Andrew about teaching.

T​he only (minor) quibble I have is that the final lab is a bigger jump in difficulty than I was expecting, but there is definitely enough help provided within the lab itself that it's still doable.

This is an excellent course on supervised lachine learning. The programming assignments are in python.

I have completed the previous machine learning course (programming in Octave ) by Andrew Ng hence I was comfortable with the concepts.

I was new to python and Jupeter notebook. Python implementation part (programming and explanation) is very friendly. I sincerely thank the mentor for immediate help on my problems in programming.

I comleted all assignments succesfully. But the strength of this course is also in the programming material given.This material is comprehensive, very rich and extremely useful. I need to go through in detail. I feel going through course material will help me to be comfortable in reading, writing, developing python programs for ML applications.

A big thanks to Professor Andrew Ng, Mentors and the deep learning community.

I strongly recommend the course for everyone interested in AI/ML.

This is really a fantastic course as it provides hands-on machine learning experience, but also a lot of intuition as Andrew is so brilliant at explaining complex concepts in very simple and understandable language and visualizations.

It is very friendly to non-math students as well as high school math such as basic linear algebra and calculus may suffice to get a lot of intuition yet without being too overwhelmed by the formality of math.

I also really like the structure of the course, and I now understand very well concepts such as the loss of a single data entry, aggregating losses into an overall cost function, and using the gradient descent algorithm to minimize the cost function to find optimal parameters for learning a curve that fits the input data.

This is an excellent introduction - I love Andrew Ng's courses! - it is exceptionally clear in defining terms, concepts and algorithms and steers a very sensibke course with respect to the associated mathematics making it the perfect first course in Machine Learning. Moving the course to python was essential and it is good to see a lot of example notebooks with supplementary material in. I'd recommend students look at Geron's OReilly Book (Hands On Machine Learning ...) afterwards to see more coding examples in the book and associated github repo. One gripe was that you didn't make students do vectorised code for the two programming asignments. I commented out the example code in week 3 asignment and substituted vector code (which runs fast).

Hi,

I already took Andrew NGs "Machine Learning" course a few years ago.

Taking it again (in Python this time) was a great refresher !

Although I understand the need to make the course more accessible I feel like the math was oversimplified at times( standard deviation, probabilities, core math functions).

Moreover I think the course should have covered EDA and feature selection before introducing supervised algorithms.

Finally, I was a bit dissapointed by the scikit learn optionnal lab, I expected more.

Still, I feel like this is the best introduction to machine learning there is.

There is a great balance between theory and practice and I like how Andrew calls upon our intuition.

This is why I give this course 5 stars.

Terrific !!! This is an excellant course that give you in-depth intuition behind the famous regression and classification algorithms. Though most of these algorithms are now readily available in scikit learn, however it's better to understand them before using them blindly. This could also help you to reate an algorithm of your own.

None the less the exercise are good and the jupyter labs are exceptionals with interactive examples.

I would highly recommend this course specialization to anyone who wants to start their machine learning journey.

Respected Andrew Ng and his team are incredible. I am really grateful and learn a lot of good things from this course.

This course is well taught, its both an upgrade and downgrade to the old version of the course. Improvements are that you will be using Python rather than MATLAB/ Octave, smoother video quality and ease of understanding, with smaller bitesize chunks of videos that the longer videos in the old version with quizzes in between taught section rather than at the very end of a week. Only dwonside would be is that less mathematics is needed and doesn't go into the detail that the old course would have, however the course was designed for people with a less mathematical background. Honest;y loved the course so far and cannot wait to dive into the next two courses.

The mathematics is explained in detail, it is true you don't need much mathematical knowledge, pre-calculus knowledge is just fine and helps with intuition, otherwise, you are taken care of with everything explained in detail. The quizzes are very helpful in checking whether you understood the concepts. I loved the labs because there was a lab for each section which gave me hands-on practice, seeing exactly what was going on and learning to apply the concepts. I am extremely grateful for the opportunity to have all this knowledge available to me across the world, this is a great course, and I loved it.

Very well structured course with great explanations in the appropriate pace. The maths are discribed clearly and the connection between algebra and algorithms (Machine Learning) becomes and easy process.

The assignments are in the indermediate level and the student should understand the theory/maths to complete them with 100% grade. They are all explained in the lectures videos but you need to think before you submit them.

Overall, is an upgrade of the previous course that is adjusted on Python and Jupyter Notebooks. 5/5 stars.

While I expected this to be simple Python refresher on the originally taken old course with MatLab/Ocatve, carefully reading into the code before executing it helped to conceptualie what I amd doing more. Also I really appreciate the interative demos, and especially those of gradient descent - they really add so much more to building your intuition -- make sure to click in the horizontal direction more anf more to the right and think why the results are changing in the way they do...

Thanks for this experience!

I really adore listening to Mr. Andrew Ng, especially when he tells something along this line, "You don't need to worry about that." This course is very beneficial for me, because I am training to become a Machine Learning Practitioner. What I learned from this course will really be what my job will be. Thank you Mr. Andrew and to the whole team who developed this course. You have developed in me the intuition I need to be an equipped and responsible Machine Learning Practitioner in the future.

One of the best course on Machine Learning on the Internet. The teaching methodology of the instructor is amazing. He is humble and explains everything from maths to the implementation of models from scratch and also with the help of libraries. The lectures are so well organized and well planned. This course actually set up the fundamentals of machine learning very strong with all of the insights + maths + coding. It actually helped me a lot about what actually ML can do in reality.

E​xcellent course! Clear insight given by Andrew on complex concepts using simple examples. Alternative way of teaching this course would be getting into linear algebra and calculus, but then then learners would have missed practical aspects of this course. I liked the fact that the focus is on practical applications. Optional labs were very useful. They gave crisp demonstrations of concepts covered in the videos. As a beginner with python, I learnt a great deal of pythons as well.

El curso es muy claro y bien dictado. Es me jor que el curso de achine Learning que estoy tomando también con NG. Recomendaría unos ejemplos mas trabajados y un curso previo de Python, pues esta es la parte que me costó mas trabajo. Aunque los ejemplos son en Jupieter, hay mucho software oculto que uno no puede entender y analizar. Lástima que por los costos no pueda continuar con los demás cursos, por eso quedo a la espera de la ayuda financiera.

Gracias

Carlos J. Gorricho

I completed the Machine Learning course a few years ago and wanted to refresh my knowledge with this new speciality course. I think that this course was far superior. The use of Python in conjunction with the Jupyter notebooks isa big improvement over the Octave/Matlab usage in the first course. And my congratulations to the builders of the optional labs. They are extremely detailed and appear to have taken a lot of work in design and implementation.

Thank you.

Tom Mone

This is the best course for revising the fundamentals of Supervised ML. I enjoyed the thorough explanation by the course instructor Andre Ng. Many topics covered in this course such as Regularization, Overfitting, and Gradient Descent are building blocks for Deep learning(NN). The practical approach for the implementation along with the conventions stated by Andrew makes this course the best among all. Overall, I enjoyed a lot learning the first part of this course.

The course covers the basics, which help understand the concepts behind regression and classification.

The labs are in python, which makes it easier to follow.

Programming the cost function and many other functions manually instead of just using the library helps understand the concept.

Finally, the instructor Andrew NG is calm and composed in explaining the complex concepts and making them easy to understand.

I thank the entire team for coming up with this course.

Excellent course. Thanks to organizers, managers, lecturers, developers, etc. It was very interesting, very funny and very helpful. I am a senior data scientist and delivering ML lectures in different universities. In spite of that, I have found this elementary course quite valuable. It is always recommended to rethink known topics and ideas and see how other specialists are delivering those important concepts.

T​hought it was great and felt it was much more beginner-friendly than the previous course. The programming aspect of it can be tricky if you've never had programming experience, so I highly recommend you learn the basics of python (variables, for-loops, functions, etc.) before taking the course. If you have some brief background in ML and programming you should be able to finish this course relatively quickly.

It's a great course for the beginners in the area of machine learning. You should have some Python basics. Optional labs are great and you can learn a lot alone, if you have desire to investigate a bit their implementations. Final graded labs are also not so hard. I would maybe add more question to quizes and more exercises, examples and datasets. But it's, as said, still a really great course and thank you!

With some Python experience, this was super hands on and easy to understand. I came into this course without a strong knowledge of how to decodify math algorithms, and with Andrew's super clear explanations and the super hands on optional labs, I found myself able to see how the alorithm was changing through visual graphs, and become able to apply the machine learning mathematical algorithms into code.