Opiniones y comentarios de aprendices correspondientes a Deep Learning in Computer Vision por parte de HSE University

Very poor instructions in course, can' even understand a thing.

1.Tutor does not explain clearly what exactly is happening, just reads the formula on screen as it is.

2.What is kernel and why we use it. Also what happens exactly when we convolutionalize,image matrix. Nothingggg is explained.

3. How one should solve the maths in graded assignments, when can't even understand the math behind the technique.

4. Speaking accent of tutor is very bad, can't understand if captions are not enabled.

I have taken many courses on coursera, which were very informative and in-depth explanation, but this course is just like National research university has nothing to do with if students are actually learning or not. It is more like earning money only.

There are several other cons which I have not mentioned.Very very disappointed.

Don't just read what's written on the projector. Try explaining it. And explain with code.

Pros:

- excellent and challenging exercises

- relevant topics

Cons:

- poor feedback and course management

- lectures that do not teach you anything, it is more a taxonomy of what exists out there than explanation of anything

Pros:

-thorough course material

-ambitious and interesting assignments

Cons:

-there isn't ANY assistance from the instructors or the TA. If you check TAs then you'll see that all of them written 0 posts on the forum.

-most assignments are severely underspecified. This makes them much harder than they should be, and it makes students spend hours on minutiae of preprocessing instead of more important stuff.

Worst presentation ever! Hard to follow, he reads from a text with no explanation. The worst! I also have to read the transcript to understand what he is saying.

Lecturer had a difficult accent to understand. Briefly put up formulas and graphs but didn't explain them. No examples were worked through yet the quiz required this to be done. To understand this subject I will have to go away and study this from another source before I can understand it.

LECTURES

-> Instructors read scripts during the whole course

-> No lecture slide (seriously ? thanks a lot for the effort)

-> Lectures are instructive even if you already know a bit about computer vision

-> Some topics and technical points are unclear, and quickly done with

ASSIGNMENTS

-> ALL assignments are peer graded, meaning you never know if you will get a grade before deadline

-> Forums and discussions are deserted, you won't find any help there

-> Instructors provide very little explanations and directions

-> Some assignments can be pretty tough to complete if you are not strongly familiar with CV

-> Some assignments are unrelated to lectures (Week1 Alignment, Week5 GAN)

Overall the course is really interesting and instructive, but it feels like it was designed in a rush.

You can look elsewhere for a good course about computer vision.

Slides are not available, only the videos. Quiz questions are very unclear and ambiguous. Assignments have almost no direction and require many hours of commitment. Requires a significant amount of prior knowledge before taking this course. Would not recommend. Andrew Ng's deeplearning.ai specialization is much better.

Course is very raw. Number of mistakes, missing lectures, missing subtitles, missing slides.Quality of lectures, especially in the starting weeks. From my point of view, teachers just read subtitles under the camera, making strange pauses, which ruins understanding.Overwhelming tasks (taking more than person's 20h each) with too few help.And no answers / support from the instructors during the course at all. Which is most disappointing, because other things can be fixed / tuned somehow.I will argue anybody from taking this course, until it's reworked and staff started to participate in the forums.

Excellent course! Quiz questions are conceptual and challenging and assignments are pretty rigorous and 100% practical application oriented.

Some good topics and interesting homeworks, however, the lectures are rushed through without much explanation or examples, and there are many frustrating mistakes that belie the fact that this course is apparently not maintained and there is no support for students on the forums. Could be really good, but it's not there yet.

Great content! Great assignments ideas. Problems with graders. Quizes and especially assignments, are loosely related to the lecture materials. The ability of authors to explain things is pretty week (it is not even close to Andrew Ng skills). The efforts necessary to complete assignments are ridiculously underestimated. You will need 10-20 hours or even more for each assignment, instead of 2hr specified. Total experience is very contradictory. I chose between 3 and 4. 3 is too low mark for these unique content, and creative assignment ideas, and I hope the assignment format could be improved in the feature, with not so many efforts (some tips, and intermediate assertions could help a lot). As for now, I could not recommend this course for wide audience, only for the most "desperate" students.

The instructor is not explaining anything. Just reading off the slides.

Lectures are good (mostly). But home assignments are very bad. Some links for datasets are broken. Admins did not fix it after a month (I found a message about it on the forum)!!! There are a lot of mistakes in a code. So, I spent a huge amount of time just for fixing bugs. It was very painful.

I think, administrators should read comments sometimes and update repos.

Some home assignments are very cool. But, mostly it is not possible to run it on a low-class laptop.

Moreover, there are a lot of mistakes in quizzes that was not fixed for a long time! So, people tried a grid search for answers. I definetely recommend to administration start fixing this bugs. This course is very good, but all this bugs discourage any desire to pass it.

The worst course I ever had on Coursera. To be honest I do thing, that it is ruining the whole specialization. None of the instructors is caring and it is seen everywhere. The sum of posts for all instructors and mentors on forum is 0, which means you can ask, but they do wont answer. All the assignment has to be reviewed by peers, because it seems that none of the author even cared about making the grader (even I do thing it is possible for most of the assignments). Quizzes are also somehow put together, but you wont get much from them.

Instructors are only reading and not explaining, so there is a lot of moments, where instructor just stop saying something because he is waiting for the text. Instructors even do not care about putting the slides for the course.

The lectures don't explain anything, they are like motivation for the topic. Programming assignments have lots of bugs and they come down to tuning hyperparameters to get a passing grade.

not managed well, not much effort from the course organizers

I am not sure how to rate this course exactly.

1, It is difficult to study. mainly because it just list most tech in computer vision but without detailed explanation.it is not better than reading the origin papers.

2, Some quizz are very confusing, Actually adding a little explanation will make the quizz more clear.

3, The assignments are not uniform. it did not give a uniform environments. you need setup the env by yourself.

4, But when you finished the course, you will learn a lot things.

the descriptions of homeworks are sometimes ambiguous, students may spend too much time trying to understand. Also, I strongly recommend teachers can interactive with the PPT when lecturing, instead of reading the draft directly.

Very bad course in quality of performance. The authors created it, but did not support it. Multiple errors, lack of help on the forum and the absence of correction of obvious problems (link to download the GAN dataset)

No relation of videos and tasks

By far, this course has the most challenging assignments in the specialization with an inherent relationship among several weeks project. Took hours and hours to beat the passing score. But it did help achieve a deeper understanding over the problems and corresponding solutions. Anyway, it is just an online course, spending so many efforts may not fit for everybody.

Overall:

This is an advanced course. Most of the time it would be difficult. In general, this course looks like the whole educational system in Russia - the given task is much harder then you can solve right now. But if you had a strong motivation you became much smarter. Support for this course doesn't even exist, but sometimes it is very useful to read the discussion to week practice.

Lectures and quizzes:

Teachers try to give a LOT of information in a very short time. But this theme is really not as simple to be told quickly. So as a result, you can understand just nothing:) Additionally, a heavy accent and kind of advanced vocabulary can make you feel frustrated.

BUT you can look at this that some keywords to google it was given) So generally, to understand I spend a lot of hours of reading about approaches in Computer Vision, mentioned here. The course was very useful from this point of view.

Quizzes sometimes look like crap. For some tasks, I still don't know to answer correctly.

Practice:

This is a part which could you a reason for you to finish this course. Tasks here really hard and requires from you a lot of time (about 20h, including the time to understand the task itself). I highly recommend all this practice carefully, it could be helpful for you in the future;) As a result, I knew a lot of tricks on how to prepare a model for find, recognize and even create faces.

This course is supossed to be a profundization about computer vision in the specialization. But I feel like this don't contribute many more than introduction to deep learning course in week 3 and 4. Notebooks are more or less the same, and even the notebooks in the introduction course are better structured for a virtual platform than in this course. Download full datasets is very annoying and unnecesary, you can spend all the estimated time only downloading datasets, this is another aspect better managed for the introduction course in the specialization. Lectures are a cute introduction but not a profundization. It seems like organizers of this course work apart from the organizers of the other courses in the specialization and didn't take enough time estructuring the course content and the assignments (All of that are peer graded and the evaluation using the grader has a better feed).

I'm sorry but I have taken the other courses in the specialization and this is the one that leave much to be desired.

Some good and challenging assignements but overall the material is delivered poorly. Slides/Videos are rushed through, quizzes are messy and there is barely any help given. I was excited to do the course as a followup from deeplearning.ai but it ended being a chore. I still learned useful skills and concepts overall but be warned.