Introduction to Deep Learning に戻る

4.6

星

1,296件の評価

•

290件のレビュー

The goal of this course is to give learners basic understanding of modern neural networks and their applications in computer vision and natural language understanding. The course starts with a recap of linear models and discussion of stochastic optimization methods that are crucial for training deep neural networks. Learners will study all popular building blocks of neural networks including fully connected layers, convolutional and recurrent layers.
Learners will use these building blocks to define complex modern architectures in TensorFlow and Keras frameworks. In the course project learner will implement deep neural network for the task of image captioning which solves the problem of giving a text description for an input image.
The prerequisites for this course are:
1) Basic knowledge of Python.
2) Basic linear algebra and probability.
Please note that this is an advanced course and we assume basic knowledge of machine learning. You should understand:
1) Linear regression: mean squared error, analytical solution.
2) Logistic regression: model, cross-entropy loss, class probability estimation.
3) Gradient descent for linear models. Derivatives of MSE and cross-entropy loss functions.
4) The problem of overfitting.
5) Regularization for linear models.
Do you have technical problems? Write to us: coursera@hse.ru...

Sep 20, 2019

one of the excellent courses in deep learning. As stated its advanced and enjoyed a lot in solving the assignments. looking forward for more such courses especially in Natural language processing

Jun 02, 2019

one of the best courses I have attended. clear explanation, clear examples, amazing quizzes & Programming Assignment this course is advanced level, don't enroll it if you are a new starter.

フィルター：

by Daniel

•Jul 19, 2018

I have completed other ML courses at Coursera, this is one which I will NOT continue. The lectures, the assignments and the grading are all riddled with mistakes. Alone that is not a problem -- however the instructors have failed to make corrections.

I am willing to push through material containing errors, however to find an error posted in the forums, a response from the instructors stating "fixing it" ... and then six months later no changes, is too much for me.

by Alexey S

•Dec 28, 2017

You should only take this class, if you already know 90-95% of what it of supposed to teach.

In this case, you might extract something useful from it.

Otherwise, it will cause a lot of frustration - the course is terrible from a learning standpoint.

by Dmitry

•Jan 18, 2019

Alexander Panin has ruined this course with his pronunciation

P.S. finished the course with honors

by Sandeep P

•Jul 31, 2018

I have started taking this course after completing Andrew Ng's Deep learning Specialization. This course is very hands-on and would be a great addition to any one interested in Machine learning. The programming assignments are harder but are rewarding in asserting the skillset.

by Marco

•Feb 11, 2018

Simply this is not a course as it's not teaching anything, but just presenting things you need to study in order to complete the exercises.

I'm on the second week and the slides and videos are simply a nightmare, can't find a better word.

The speaker is not able to speak a fluent English and you can't really build a logic sentence even from the subtitles. You get some intuition from the slides and from some comments in the forums.

I'm not struggling with the actual topics but to translate them into Tensorflow code, which is not a pre-requisite for the course.

I'm wasting a lot of time trying to understand how to do the silliest things, just because there isn't any introduction to it. I have painfully lost hours trying to understand how to do a reshape of a placeholder with not known sizes, but you are supposed to finish the exercises in 1h, while this is barely impossible if you do not know Tensorflow in advance.

It looks like that you need to know both machine learning topics and Tensorflow, which means that you don't really need this course then.

So I'm not sure what is this course about as it is not really teaching anything, topics are just presented and then you need to do your own research. It looks like a book page table: you know what the course is talking about, but you don't get any real explanation to that.

I know that the course is considered "advanced", but it does not help you in solving the exercises at all as you need to learn elsewhere how to do it.

The time you spend for the exercises is massive but generally it is not correlated with the difficulty of the topic, but more in the way the exercises are presented.

It's a bit of shame as the teachers look very competent, but it really looks as if they haven't put a good effort in this course.

by Yevgen A

•Sep 25, 2019

Great course. Even though I've done Andrew Ng's ML course twice and completed his Deep Learning Specialization, I learned a lot of new things in this Intro Course of AML specialization.

by Radishevski V

•Nov 27, 2018

The course is good enough, but lecturer Aleksendr Panin speaks too quickly and anyway with a strong accent. Fast does not mean good

by Nikita F

•May 09, 2018

Most video lectures are useless, lecturers are just reading some text from a paper/screen, some of them have english very far away from perfect. Each programming assignment is more about struggling with bugs rather than learning something about ML. Final project is a torture for students without GPU, you spent 5 hours training the model - your final loss was too low. Can you add a test/assertion for the loss function?

So far it was my worst coursera experience.

by Darya L

•Dec 14, 2018

In general the course is good, it gives you the idea of different neural networks, their usage and a bit of their inner math. The only thing I didn't really like: most programming assighnments contain large precoded parts, which are difficult to understand. For me it would be more useful, if assighnments wouldn't be so difficult, but I had to code myself.

by Anna N

•Mar 07, 2018

Lectures provide very small amount of material. There is no sense to describe topics like gradient descent in advance course. It would be much better to take just a few topics and describe them in much more details than to speak 5 min about CNN, 10 mins about RNN, etc.

Also big minus is poor English.

by Артем

•Oct 15, 2018

Just a theory, no practice at all !!!

by Lewis B

•Jul 22, 2019

Overall a very good course. Highly informative, and strikes a good balance of the application of neural networks and theoretical background. I should mention that I have studied mathematics to MSc level so I didn't find the mathematical aspects of the course challenging but this is will vary depending on your own background - previous study of multivariate calculus will help a lot.

I particularly enjoyed the tensorflow content and found this to be particularly well taught, in fact it is the best introduction that I have found for this module. Personally, I would like to have seen more of autoencoders and less of CNNs but this is probably due to my own individual area of application - time series are more relevant to me than computer vision - and it really is only a preference.

Lastly, I believe if you are going to learn something it is worth learning it properly and comprehensively. This course does that and I doubt you will find a better introduction to neural networks in that respect.

by Daniel I

•May 08, 2018

PROS: Interesting exercises

CONS: Very poorly explained. Poorly prepared exercises.

Hard to understand if you don't already know about the matter (then... Why would you need this course?)

by Vratislav H

•Nov 05, 2018

This course is one of the most difficult I have seen but at the same time it is very well structured. Lectures are understandable, one just need some support from other materials to understand a whole content, at least for me. I struggled a bit with a final project but in general, I enjoyed it a lot, I looked forward to it each week, it was challenging and achievable. I recommend it.

by YaMolekula

•Jul 27, 2019

I think it's the best intro to DL, especially thank you a lot for mathematical explanation of marix/tensor derivatives. Also implementation NN with numpy and gradient descend improvement are my fave tasks.

by Anas K

•Jun 02, 2019

one of the best courses I have attended. clear explanation, clear examples, amazing quizzes & Programming Assignment this course is advanced level, don't enroll it if you are a new starter.

by Rahul K

•Mar 01, 2019

Really Great course. I would recommend everyone to take this course but after having some "basic knowledge" of Machine Learning, Deep Learning, CNN, RNN and programming in python.

by Erik G

•Apr 13, 2019

This course gives a great overview of what can be done with DNNs. Topics are well chosen, clearly presented, and a good level of difficulty.

by Oleg O

•Dec 01, 2018

Useful course, whereas it is not always clear how to complete homeassignments

by AJIT R

•Feb 28, 2019

This thing is AMAZING. This thing iS no "INTRODUCTION" - IT IS "ADVANCED".

by Andres V

•Nov 23, 2018

nice really hard course.

by Meng Z

•Jun 08, 2019

This is a good course, though the instructors failed to keep their pace. If possible, I hope the course updates along TensorFlow 2.0 and provides more readings. As mentioned by other students, we don't want to watch videos on gradient descent again and again. I hope the instructors save time to talk more about some state of the art models and more about TensorFlow, links to good readings, and maybe more exercises on gradient descent and other fundamental techniques.

by Abhishek S

•Jun 12, 2018

It's good overall.But if you are a beginner ,this course might be very advanced for you.

by Федоров Ф

•Nov 18, 2018

i think that the explanations and examples in the notebooks was not always sufficient

by Igor B

•Feb 08, 2019

The course indeed gives an introduction to deep learning, but the practical part is discouraging since the "deep learning" part of practical assignments is usually given rather than asked to develop individually.