A Review of MIT’s Probability and Statistics Course on edX
https://courses.edx.org/courses/course-v1:MITx+6.431x+2T2019/course/
Disclaimer : This is my first ever article, so please be patient with me!
Let me give a little background about myself. I am a BS Computer Engineering graduate, and I’ve actually already taken a similar course, Probability and Statistics for Electrical and Electronics Engineers, during my undergraduate period.
You might be asking why would I take another similar course…
Well… Unfortunately, I did not really understand what was happening during my undergraduate. I couldn’t follow the professor during his lectures, and there would be rare occasions wherein sleep transported me into another world.
Fast forward to now, as I aspire to become a Machine Learning Engineer, I believe that understanding the basics and having a solid foundation when it comes to statistics is essential, as Machine Learning is just glorified statistical learning. That is why I thought it was well worth it to retake a similar course. And, hopefully, you would be able to make the decision through the aid of this article.
Now, let’s get started!
I enrolled in the course for free, which started in May and in September. By free, I mean that the platform edX allows you audit the course, allowing you to watch its videos and answer exercises; but you would not be able to get a certificate at the end.
The course was hosted in edX which was alright; but sometime in the middle of the course, MIT decided to implement new edX rules wherein the second and third exam were made exclusive to people who paid for the course. It was quite disappointing to me, since I wanted to see a passing score at least, even if it wasn’t going to be certified; nevertheless, I was able to learn a lot, and I am very thankful to MIT. Okay, enough of the chit-chat! Let’s get into the meaty part of this article.
As I did not have a solid foundation in Statistics, I allotted 7 to 8 hours for it on a weekly basis. However, if you had already taken a similar course and managed to pass with flying colours, maybe you can do away with half the amount!
The course, like any other MOOCs, rely on videos to educate the viewers about the subject matter. It somewhat follows a routine as stated below.
1. Watch a few videos.
2. Answer the exercise.
3. Repeat steps 1 and 2 until you hit the end of the unit.
4. Answer the problem set.
5. Answer the exams once they’re released.
6. Finish the course!
The videos usually involves a blank page at the start, and eventually, a bunch of formulas at the end. A person, your lecturer, would be writing derivations and giving examples throughout the video; and at the same time, guiding you with his words and insights.
For the exercises portion, normally, you would just have to pay attention to the video that precedes it. However, for the problem sets, that’s when you have to think outside the box and similarly for the exams.
It would usually take me around 15 to 30 minutes to solve the exercise, 4 to 5 hours for the problem sets, and half a day to a whole for the exams.
I summed this course up to three criteria as you can see below, and hopefully these would be able to guide you in making the choice of whether to enrol in this course or not.
Course Content : 9/10
The course covered topics such as Combinatorics, Counting, PDFs, CDFs to Random Variables, Bayesian inference and many more; the syllabus can be found on the site. From what I’ve learned, I could say that this course has covered most if not all of the fundamentals of statistics. In my opinion, as long as you devote time to actually understand the exercises, problem sets and exams; it would prove very helpful in the long run.
I did feel, though, that this course had too many topics to teach in just a span of around 5 months or so. Honestly, I could not keep up with the deadlines if I did try to finish all the videos; so I just kept to the bare essentials and hoped for the best when it came to answering the exams.
I think the most important lesson that can be gained from this course is to be patient in understanding complex concepts that would require more than a night’s worth of studying to fully grasp. The Q&A part, which was (usually located at the lower part of the page, was very helpful in answering the exercises; as the staff or the community offered tips to help you get past the exercises.
Time Scheduling: 6/10
I can rightfully say that I am no genius when it comes to solving statistical problems; and that is why I had to cram the exams by burning the midnight oil and finally be able to sleep at 3 or 4 AM.
I work at a 9 to 6 job, and I also played badminton during the night; so I had a difficult yet fun experience, in retrospect, balancing my work life, my night life, and my homework.
The most helpful thing the platform offers is the ability to download the videos, so you do not need constant access to the internet in order to progress through the units. Make sure to use that to your advantage! I actually had to use my breaks during work hours to watch the videos on my phone and answer the exercises whenever I could.
Allocating enough time for this course is a must.
Trust me, if you don’t; then, you will have a hard time going through the future lessons, as your lack of understanding would cascade onto the future lessons. My advice is to persevere through the exercises, and give yourself enough time; otherwise you will not make it to the end of the course.
Will I recommend this course to another aspiring ML Engineer : ?
Do I want another innocent person to endure all of the sleepless nights I’ve went through?
Absolutely.
This course, apart from teaching you the technical lessons it should contain, will also teach you life lessons as well.
If you really want to achieve something later in life, you will have to commit your heart and soul to reach it.
This means sacrifice is necessary.
We’re talking about time and not human sacrifices, so don’t worry. There were nights when I had to give up badminton just so that I could finish the drills on time, and believe me when I say that this was a challenge in itself.
P.S. I really love badminton.
The second important lesson it taught me was to manage my time well.
When you’re only free for a certain number of hours per week, you would have to weigh in your priorities as to which would get the huge chunk of your time. I thought that this course was extremely important in my growth as a ML engineer, and that’s why I decided to put in the hours for it.
The third and final lesson I’ve learned is probably to be patient with yourself.
It can be very frustrating when you can’t understand one problem given that you’ve already spent countless hours on the videos. More so when some of your virtual classmates seem to understand every particular question you can’t comprehend. This can be very discouraging and might prevent you from reaching the finish line; but be patient.
The course is not asking you to race against your classmates, but rather for you to portion enough time so that you will get to the finish line in the end.
In conclusion, the main takeaway is to understand the fundamentals and building blocks of statistical learning. What is more important, however, are not those technical lessons; but the amount of discipline you will enforce on yourself. And just maybe, you would be able to carry on with you that same discipline into other difficulties as well.
Credits to my girl, Chynna, for the initial proofreading.
