What expectation of optional labs as a student?

In some optional labs, there is much new knowledge, and syntax, especially custom imported libraries. Those make me not understand much, or understand about 50 - 60%.

Are there any problems if I just go through, and study the next videos without understanding a particular lab? What expectation of optional labs as a student?

Hello @Nhat_Minh,

At the end of the day, what your expectation is matters the most. You will have to decide it, but I can share some personal opinions. There are 4 aspects that I think is relevant to the course:

  1. Python. But MLS is not for teaching Python. I won’t expect to learn Python here, but along the way I will take away whatever it looks interesting to me.

  2. ML Concept. Demonstrating ML concept taught in the videos is I think a major target for labs. The demonstration is 2-way because jupyter notebook is an interactive environment. One way is to see what is presented to me by default; another way is for me to adjust any adjustable parameters or code lines to discover different angles of the concept. The first way is minimum, the second way is premium because I am going to challenge myself my understanding of the concepts in those different scenarios created by adjusting the parameters or code lines.

  3. Translating ML concepts to code. Here, I don’t just challenge my understanding of ML concepts, but also see if I can understand why the codes are coded that way. Labs provide a rich description of some major implementations coming next. If the description talks about a maths formula, am I able to understand how the formula is coded. If the description talks about certain if-else controls, am I able to understand how the variables are manipulated to achieve the control. For us to break into AI, being able to code what we think is a must. Labs give you the description and the ready code for comparison.

  4. Debugging. We don’t need to code in optional lab but we do in assignment. Sometimes we get errors there and don’t know what to do. The good news is sometimes optionals lab actually demonstrated quite similar implementation. It gives us a baseline working code to compare with our erroneous version. Am I able to make use of the working code to debug my non-working code? Am I able to trigger an error that is fully under my control? Writing working code is critical, but debugging non-working code is as critical because the latter usually happens more often than the former.

Finally, in setting my expectation, I think there are some relevant questions to ask myself:

What is my current status? student / working in AI / working in non-AI but looking for a change
What is my current goal - the achievement I am pursuing and hopefully MLS can give it a push ?

With those answers, I will adjust the proportion or effort I spend in the 4 areas, and that will form the shape of my expectation to the course.



Thank you so much​:heart: Raymond

You are welcome Nhat!