Thank you so much to the deeplearning.ai team for the course!

Thanks to your work, you have been able to immerse yourself in the world of medical applications for ML.

I will highlight the following points as a criticism:

* imho, mediocre code quality: vectorization is ignored in many places without proper reasons; insufficient code encapsulation: beginners can learn bad things

* many programming tasks are more like solving riddles and require more time to solve them than contain useful information

I wish you good luck in educational field and look forward to new courses from your team.

This is the worst course in all my experience on Coursera. The material has been prepared with evident hurry and superficiality but the more disgusting thing is the total absence of assistance and support. There are evident problems with the grader, reported by several users, and no one has been giving an answer since the course startup. If this is the quality of service provided by Coursera, there is non hope that I'll engage a new course in the future.

This course was quite disappointing. Although I expected more, given the track record of deeplearning.ai, this course felt rushed, and more of a practice in good coding rather than giving any knowledge or advanced intuition into the applications of AI in medical treatment. The first assignment in particular felt like it was mostly testing the student's ability to use fancy python package syntax than anything else. Overall, I wouldn't recommend this course at the point of competition (November 2020). I think students are far better learning the AI methods described in this course on their own, and only then learn their applications in medicine if this is required for their work. At the end of the day, data is data no matter what we're looking at. This is not to say that this course cannot improve in the future. To be fair, I would actually hope it does, to keep the great standard that deeplearning.ai has gotten us used to over the years.

Most pathetic course done so far !

The whole specialization was well-designed in such a way that even an engineering graduate with very less knowledge about statistics, probability, and also machine learning could understand what was taught. All the videos were short and crisp to the point and were quite explanatory. The assignments were really good and sometimes challenging which really made me think about the solutions. Except for the week 1 quiz that was repetition from last course others were really good. The entire experience as a CS undergrad new to AI in the field of medicine was great. Hope we get to experience similar courses in due course of time. Thank you!!!

Excellent course and the specialization. I feel like I participated in a research project. Learned much, and have cool notebooks to revisit at depth.

This was a great course for learning how to understand what your model in doing. Being a data scientist in industry, this is really helpful for error analysis. Thank you to the instructors for creating this entire medicine specialization. It is one of the most practical AI courses that is out there thus far.

This is a fantastic course. I have learned so many new things about the application of AI and deep learning in medical treatment. Using the BERT model to answer questions from medical report was surreal.

I found this course very interesting and completely worth it. It made me notice how the subjects splits should be performed and I also particularly enjoyed the part in which we apply GradCAM.

Learned a lot about interpretations of both machine learning and deep learning models. Introduction to basic NLP techniques was a great start too. The overall course is really good.

While it is understandable that not many techniques are covered duo to time restriction, it will be better to include a list of reference at the end to let students know there are also other techniques to solve the problem (e.g. different methods for AI interpretability)

Some of the content in the slides are wrong and very confusing. While a textbox correcting the error is appreciated, the better way should be to correct them directly in the video by editing the video or simply re-record it, given the length of each video is quite short.

The theory is good. Like, it was the first Ehealthcare course, where someone told about SHAP and different types of missing data. But programming tasks are huge and boring. As this course is not about the deep knowledge, then why are you making us implement cross-validation in raw python when it is already implemented in standard packages? Better teach how to use it from there, IMHO. And the original code full of things like `from utils import *`, global variables, etc.

Provides a good introduction to RCT, evaluation of treatment effects, using BERT for question answering, label extraction from medical reports and also interpretation of ML models and deep learning CNN models. The videos provide a good understanding and the notebooks in the exercises and the assignments help in applying the learning, so that the learning is reinforced!

some parts of the labs not mentioned directly or sometime even indirectly in the videos, previous courses had rich lectures however this one require to add some missing parts into the videos to make sure all labs and assignment covered in a way.

Thank you so much for the quality of those course..deeplearning.ai please keep lunching such excellent courses en AI : Ai and computer Vision, AI for industry, Reinforcement Learning,etc.

A very important course. Only 4 points, because at many points I had a feeling that many things were abstracted away, and am not sure whether I'd be able to replicate them on my own.

This is more of a course on pandas dataframes and various data preprocessing and formatting tasks than really understanding the functionality of scikit-learn and tensoflow as applied to medical research. The instructors should spend more time on presenting the AI tools proper than dealing with the minutiae of of how to manipulate the data. I'm not saying that it is not important to structure and manipulate the data, but at least equal (in fact more) time should be dedicated to understanding and using the AI tools. Disappointing!

This name of this course is in my opinion wrong. It is much more about model tuning and evaluation than medical treatment in itself. Though it gives very useful and numerous techniques for this purpose, the assignments are quite boring, especially implementing the gridsearch in week 1 (quite out of the scope of the course, and much too long) or copying/pasting Keras functions in week 3.

The videos are however once again very clear and interesting.

course must be updated to latest version of TensorFlow. please shows us how to train your model or add exercise to train the model instead of loading it. you can switch to google Colab for more processing power the way they did it into TensorFlow dev. certification course.

The lecture videos are great, but the Jupyter labs are entirely convoluted.