Avis et commentaires pour d'étudiants pour Advanced Machine Learning and Signal Processing par IBM

It is supposed to be an advanced course but there is almost no advanced topic in this course and it is just a shallow overview of machine learning with the Pyspark. SystemML explanation was very vague and incomplete.

The assignments are too easy, the level of coding required is not very challenging, it's just a fill-in-the blanks exercise, I don't know if I could actually do any of these things on my own with a new data set.

Excellent brief math lectures by Manchev. The course materials do MOST of the programming for you and so you only get a light exposure to the Apache Spark API -- insufficient to develop real proficiency.

Overall very useful material covered however I was disappointed that some key concepts such as Baynesian inference and PCA were not well explained. I supplemented most of that material from Youtube.

The learner needs to do more by his own. I think the course should follow up on the teaching style from the IBM specialization of Data Science. The teachers are good at replies.

Assignments are too easy, and not cover every lecture that I have learned for advanced ML, also some of the lectures are quite short.

The weekly assignments are quite bad, and the course name ,advanced ML and signal processing, is misleading

Not enough programming assignments and the ones in this course are too easy for the "advanced" course

The assignments are way too easy. Not very challenging for a course with 'advanced' in its title.

Some videos with low quality so it was hard to read the code and lack of pratical exercicies

The course might need some updating, it does give a lot of information about many things.

Not too sure about the application of digital signal processing in real world problems

A tad too much IoT for my professional interests (was part of larger "Specialization")

Some explanations were good but that was not enough for covering machine learning

Hard to follow at times... found a lot of assistance in discussion forum

The spark usage is very limited. Assignments could be more challenging.

It should be useful to introduce more practical exercises

well explained, programming assignments are worthless.

It's a little bit comlicated develop the activitites

assignments should be more challenging

I honestly expected so much better from IBM.

The idea of the course is great, it covers interesting topics about distributed machine learning, how to perform it with huge amount of data and how to solve scalability problems.

The idea is great, but it's really poorly executed. The course lacks of a real structure, connection between lessons and real explanations. The only lesson well structured and well organized are the ones held by Prof. Nicolay Manchev, they're clear, well structured and splitted into theory + practice.

The lesson about SystemMl are held by another professor (not mentioned in the instructor list) and they're almost impossible to understand, and i'm not talking about the content itself (which is still pretty poorly explained) but the pronunciation, it's really really hard to understand what he's saying.

Another real problem about this course is that seems more "marketing" focused that "learning" focused. Having followed also the previous course of the specialization, you can clearly tell that some explanation are more marketing than a real explanation. I understand the point of view of IBM, it's a company, they're interested in making money and marketing, and there's nothing wrong with it. But if a student finds himself annoyed because of this, it starts to be a problem. I don't care about your amazing offers ecc... i want to learn how to use and when to use those tool. Eventually i'll start using on my own your platform, i don't need constant remainders about how beautiful it is.

Last thing last, right now seems that IBM decided to completely remove the free basic plan on their platform (happened just today, and i had some problems finishing my last programming assignment). Doing so, they literally removed the possibility to test and learn on their platform, since you're limited by the monthly credits they give you.

This is pretty funny because the goal of the course (beside teching to students) is to promote IBM and their platform, and after this course and the removal of the free envirorment from Watson studio, i completely moved to Google and Kaggle, pretty ironic.

I was initially excited about pyspark and SystemML but It's a very gentle introduction.

The assignements were way too simple machine learning wise. On one of them you just had to call the classifier. One word + (). Really?!

On the positive side, PCA, FFT and Wavelets were very well explained.

Quite an unbalanced course. Some material is very primitive, other is quite complex compared to the rest. Lab assignments could have been more elaborate. Even though I learned quite a bit, my expectations were higher for an advanced course offered by IBM

Course was hard to complete due to the version problems. Instructors should update material so that the course can be done with newest versions of programs.

I haven't got any time for doing the course, as I joined the course 1 week ago.