Comentários e feedback de alunos de Processamento da Linguagem Natural da instituição Universidade HSE

The quality of this course is good, but I still feel that the algorithm introduced is basic. It is better to combine the leading edge algorithms. Actually, I finished it very early, but I stuck in the fifth week of the dialogue robot job for a long time. I feel that this part of the time is not very worthwhile, because time is spent on the more cumbersome model deployment instead of focusing on Algorithm optimization, of course, model deployment is also very important, this is also a test for me, so I give 4 stars.

Deadline passed, no review. I submitted my Honors assignment 8 hours before the submission deadline, exactly when I was supposed to in order to ensure a peer review. I performed reviews on a number of other students' works, as required. Yet my work remained unreviewed for months, until a student from one of the next classes felt charitable.

Great introduction to NLP. But NLP is too huge topic so need to study more about this when really want to work at company.

Very interesting, qquite advanced and you need to invest time for the assignments, but worth it in my opinion

There was a lot of emphasis on mugging up formulae instead of visually explaining how they work. Even the Tensorflow exercises felt pretty convoluted. Circa 2020, there are many better ways to write the same TF code in a very legible and easy to understand way. This course needs a major refresh.

NLP ist very interesting.

BUT this course is a pain.

Material is from 2017, partly in Russian,

sometimes not working or need enormous time invest to make it running.

very technical but there is not motivation regarding the reason behind it

I didn't like the teaching way!

The course was way worse than my expectations after the first course in the series- Introduction to Deep Learning. I really liked the NLP week at the first course and the lecturer there was saying that in the actual nlp course there will be way more details and explanations like for skip gram model. However, the course almost does not offer any explanation. It mentions algorithms rather than explaining them. Like they do not explain forward backward algorithm for Baum Welch just mention it and I had to search and read about it myself. I am just writing it after I watched the doc2vec video and there were like only one or two slides of how we go from word2vec to doc2vec. Absolutely no explanation she just mentioned that those algorithms exists. Maybe I am too stupid to take this course, and everything is obvious, but for me I took this course after I finished introduction to Deep Learning because I really liked NLP explanation int he first course. Also I ask a question in the forum and nobody answers for weeks

Really really bad teaching. period.

1.in-consistent word usage. they introduce a term to you now, and use a different term meaning the same term, which has been introduced earlier, but the student will have no idea of it.

2.the way these teachers approach teaching a concept is worst I have ever seen my whole life. These teachers just read from the slides and just a sentence or two more from what is in the slide. Slides should serve only as a compliment to what a teacher talk about, not read from the slides itself.

Explanations are not in depth. Problem sets introduce knowledge not provided in the lectures. This course needs to do in-depth explanations of concepts before throwing them onto a problem set.

Full of Mathematical explanation..Nothing from code perspective

They aren't teaching codes. Just theory!

Pronunciation is terrible.

Excellent course. It is clearly advanced and the student is expected to know about many concepts/technologies before starting.

The lectures treat state-of-the-art methods for NLP using (deep) ML in a high-level fashion without going too far into the details. For this reason the course is good for learning about the availability of methods and how they work in the broad sense, but if the student wants to know more he/she should read the accompanying papers. I enjoyed listening to both lecturers since they were speaking clearly and with good structure. However, sometimes I lost them because they went over some of the mathematical topics quite quickly.

The assignments are great as well. Their level matches the expected level of an advanced ML course. Through the assignments I got a good idea of how to perform NLP in practise with the latest tools. An improvement could be giving the student more freedom in how he/she tackles the assignment (though I understand it makes reviewing more difficult).

In addition I think it would be good if there are multiple tests that check the same theory over the whole course. After each week I could make the quiz quite easily because the knowledge was fresh in my head, but if I had to make them again now it would be much more difficult.

This course is truly awesome. I tried to find a MOOC that would cover both deep learning and prior methods for NLP, but this one is the first that did. It does that very well, too - it is really comprehensive about topics that it touches - even if you know something about DL in NLP you can find something new, and the coverage of advanced techniques is outstanding.

One thing could be better though - I think that programming exercises are too easy (some parts are easily solved by plugging couple lines of code). On the other hand the last assignment is probably one of the most interesting programming tasks I've seen on courses (you need to implement a chatbot so that reviewers can actually talk with it). Also the last assignment can be expanded using more advanced methods in second task, which belongs to honors track, so you can, but you don't have to make it more challenging.

Great course ! It is a good introduction to NLP, from the basics to elaborate models. It also provides contents to further explore the topcis and other models. The assignements and quiz are challenging, but with if you have time you will pass them with the help of the staff and the forum.

The last assignement is quite hard, because it involves a lot of setup (AWS Linux instance, Docker, etc.), and more explanation materials could be useful. But it also force you to understand deeper the construction of the chatbot, making you learn more than just the NLP part.

This is a super excellent course, it does not only introduce you to natural language processing but takes all the way to be advance, in such a way that you can curate a bot straight from your mind, having no hardship thinking about the concept, i love the fact that this course does not spoon feed but gives enough information, to spark curiosity, in a way that you find yourself studying to master, through external sources, different views i loved it

This course might be one of the most challenging courses I've attended in Coursera. I've got so much practical experience. I always want to learn docker, AWS. This course 'forced' me to get to know them. I am very happy to build my first chatting bot with a lot of effort. It is worth with my time, I think. In the assignment tensorflow 1. is used which is a bit outdated. This is the only downside. Thank you teaching teams for such a great course!

This best NLP course I can find in the internet, the team is great. They explain everything from basic mathematics to implementing research papers. If you want to take this course you should have basic concepts of python, deep-learning and mathematics(even thought they explain).

I am pretty much confident to working with text data.

Thank you HSE team for preparing a good Advanced Machine Learning Specialization.

All in one place.

It is a great course having both mix of traditional ML and deep NLP approaches (the new might be better but we need to know the prevalent ways a decade ago!). Should keep the content updated with always developing field of NLP. Assignments are really useful to understand the subject and stays as a starter code for new projects you wanna start. I would like to convey my thanks to the instructors !!

Great Course on Natural Language Processing!

It requires an advanced knowledge of Statistics an Probability to be fully appreciated.

Assignments are tough but worth the effort.

On the downside:

1) the course needs to be updated with bleeding-edge technology like Transformers, Transfer Learning Techniques, BERT, GPT , ecc..

2) it uses Tensorflow 1.X version

3) The peer grading is a little bit tacky !

This course is incredibly informative and has been instrumental in allowing me to really understand what different NLP methods are doing under the hood. It is difficult and takes a lot of time, but the programming assignments will give you code that you can apply to a bunch of your own applications, and you'll have the understanding to know when each technique is appropriate!

Nice course. Thanks a lot. Great teachers/mentors, amazing exercises, I enjoyed every minute of the course. At the end of the course provided a list of papers and research materials with some trends in the field. Those people from HSE save a lot of time digging around the web to get a clear understanding of weaknesses and strong sides of modern NLP approaches.

nicely organized! amazing course. I am doing my PHD in NLP, and I had prior NLP classes in coursera, but I still can learn quite a lot knew things from this course. It gives NLP from another perspective, and it is really up-to-date with deep learning and tensor flow. Love such classes. Hope there are more classes offered from these instructors.