General course question to mentors!

other than discouraging students from learning, what is the point of making the quizzes and assignments infinitely not related to the lectures?
When less than half the folks pass a quiz, it means either the lectures are inadequate, or the questions/multiple choice are not clear. or both.
In many cases I find myself deciphering your choice of words in the multiple choice trying to figure out what you really mean.
These waste time instead of letting us move on in the course we are stuck figuring out lousy quizzes.

Dear @Ramez_Sabra,

Neither lectures are inadequate nor questions.
Quizzes are designed after lots of hard work.
Follow the lectures carefully, and you will automatically give the answer of any questions from the quiz.

I disagree with your conclusion.

The quizzes are not just a checkbox exercise, where you just look up what you already saw in a video. That would not create learning, it just verifies you watched a video.

Sometimes you have to do some thinking and maybe additional research.


Hi @Ramez_Sabra

Prof. Ng started the MOOC online courses movement can be traced back to 2011 and over the years his courses continue to develop with the aim to democratize this highly skilled level of education to a wider audience. It takes a lot of work and dedication to produce a course. If you found the course material is not up to scratch, or if it can be made clearer, then please drop a suggestion comment to the course staff, highlighting the area of concerns, location where it is found. The staff team would be most appreciated with your feedback.

There are technical words in the course lectures, quizzes and lab assignments. If you are not sure about the meaning of such technical terms, a quick search on the internet would definitely help. Alternatively, post a query on the course forum.

The forum community is very helpful and happy to answer your queries and help your understanding. So do make use of this resource.

1 Like

I remember back in 2022 when i started learning programming, where if u search in youtube will find thousands of videos or lectures of programming, but non of these doesnot make sense to me after exposing my self to a project or task i feel overwhelming and discouraged but after a while i discovered that the problem is me myself because i was just watching videos and hoping i will understand immediately build next instagram, i made a decision to build some simple projects and work out how i can improve my problem solving skills, now I’m not a pro but i can try my best, then AI and ML get my attention and similarly i tried to learning from some quick bootcamps which says learning the whole machine learning deeply learning and the likes within a week, that was not possible i may finish the video lectures but i couldnot finish one simple EDA project, then again i re-discoverred my self by knowing in order to learning from scratch and know the algorithm in very deeply way, i started taking M4ML and after i finish the linear algebra was feeling a big difference and then i continue learning the calculus and everything is in place, overall you need to figure it out your issue before you blame highly respected instructors and by the way building such like this courses takes years.

so re-discover your self and try to approach things differently and in many ways before i finish linear algebra the videos only where not enough, i have to search for an offline books and online videos which may take an alternative way of solving this problem.


1 Like

Here’s an example from a question in the quiz:
Which of the following are true about dropout?

Not definitive Q&A, more like debate points:

  • It helps reduce variance. (ok, makes sense).
  • It helps to reduce the bias. (could be yes, I can argue for this one. In fact, imagine you are sampling a population differently every time. your standard deviation is dropping, bias will surely drop as a result. I can also see the point that no direct impact on variance). Hence confusing answers.

From the very same question, here is an example of absurd wording:

  • In practice it eliminates units of each layer with a probability of keep_prob vs 1 - keep_prob!!
    What is the point of playing on the term? Frankly, both these sentences are correct. the second (1 - keep_prob) is more descriptive but without the 1 the term is still not wrong and delivers the meaning!
    Again, this is an example of not clear or definitive Q&A.
1 Like

This is my point, the lectures are not sufficient for the quizzes, we have to go out and “research” online to get more info to answer the quiz. hence the low percentage of people passing.

1 Like

Thank you and I salute the hard work going into the lectures and putting this out to the public to learn it. I just think the quizzes can be made clearer and more inline with the lectures. or add a check point after every video.

1 Like

Hi @Mohammad_Omar_Adde ,

You are absolutely right and thank you for sharing your learning journey. There are a lot to learn. The only way to get the knowledge into our head is by doing and be curious.


@Ramez_Sabra ,

Thank you for highlighting the issue that you have experienced. With your help, I am sure the course material can be made clearer and better.


Hi @Ramez_Sabra,

The dropout technique is used to reduce variance. That’s its main purpose. It does not directly address bias.

How so ? keep_prob is the probability that a unit will be retained. And the option is talking about the probability of a unit not being retained/ being eliminated.

Lets say,I will eat a chocolate has probability 0.3.

will_eat = 0.3

So tell me which of the following is correct ?

  1. Mubsi shall not eat a chocolate has probability will_eat (0.3), or
  2. Mubsi shall not eat a chocolate has probability 1 - will_eat (0.7) ?


1 Like

But remember one thing, in order education to be effective there should not be a spoon feeding way of teaching, since this is more like higher education we have to search and we learn how to research and sort out things, i know it gets difficult but after you search it shouldn’t be difficult and also the real world is a bit harder than this in my view and a lot problem solving skills is needed, thanks for commenting but also try to search and find better resources so we can share here, then those who are struggling will benefit it.


1 Like

As a student here in Deep Learning, I have to disagree with the impressions given by the fellow learner.

When I have struggled to pass a test or some particular questions it has very often been due to my lack of attention in some videos, or having rushed over the material.

The questions sometimes make us think twice, or review the material, which I don’t see as a problem, well on the contrary.

I love it when after passing the test (or failing), a couple more opportunities are given to complete the questions not answered correctly on the first try. This is a very nice challenge for a demanding comprehensive learning.


You missed my point. I am not asking for spoon feeding teaching experience. and You should not say that to me.
I am being very precise and accurate. The quizzes are very unclear and in many cases subject for debates. I can give examples on and on. Read then reply.
Here’s yet another example for the ones missing my point:

“After working on your algorithm you have to decide the next steps. Currently, human-level performance is 0.1%, training is at 2.0% and the dev set is at 2.1%. Which statement below best describes your thought process?”

Now you expect to choose one answer. The best answer. based on the question wording. But no. There are multiple answers that the learner has to use.

If one more genius here talks about spoon learning and needing to research they really should consider meditation to improve their focus skills because THAT IS NOT MY POINT.


I think, ultimately, you care, otherwise, you won’t spend your precious time voicing out.

I see your point about the last question, and I have requested the course team to take a look at the wording.

Thank you, @Ramez_Sabra.


1 Like

Hi @Ramez_Sabra,

Please be respectful of the forum and the mentors.



Thank you.
I think Prof Andrew and his team did a great job at AI learning and making it available in an easy yet complete way to everyone.

I definitely highly recommend Dr Andrew courses to everyone.

1 Like

Some of the questions are “one answer only” and others are “choose all that apply”. Usually they will say something in the wording of the question, but even in cases that they don’t explicitly say so, notice that the “choose all” ones have square check boxes, whereas the “one answer” ones use a circle. But if you didn’t notice that subtlety, you can also just try clicking a second answer and see if it unchecks your previous answer. :nerd_face:


Thank you Paul. Always looking for your helpful comments. Yes I notice that difference. I just like to keep them on the edge lol.

1 Like

Glad to hear that you were already aware of this, but it makes your previous post feel a bit disingenuous.

1 Like