Machine Learning by Andrew Ng
WARNING: This course is not recommended for people who are not familiar with linear algebra.
Course Link: https://www.coursera.org/learn/machine-learning
It was a grueling 1+ month trying to complete this course. Don’t get me wrong, it is a great course and I learnt a lot from it. The course is very informative and Andrew is a great instructor. The content though… was extremely demanding.
Dealing with differential equations and linear algebra is not something that interests me particularly. Despite that, I managed to gain valuable insights to how machine learning works and how it can be implemented.
What’s the one most important thing that I took away from the course?
DATA IS KING. Whoever has the most amount of data will be able to make the most accurate predictions using the models they have built, regardless of how bad the algorithms are. Based on this knowledge, it’s hard to see how Tesla’s advantage in developing Full Self Driving will be eroded in the near future.
That’s not to say that we can get away with bad algorithms but we definitely can get away with imperfect ones.
Pros:
1. Free course materials (no certificate provided)
2. Very detailed explanation of the concepts
3. More than enough opportunities to actually implement the concepts
Cons:
1. Have to pay for the certificate(which I didn’t as it is 79USD)
2. Likely to burnout halfway through the course
3. Nobody really uses Octave beyond the scope of the course
Notes I have taken that I probably have difficulty recalling:
https://drive.google.com/file/d/1--_RHJI56-Zm3-BE4MGdmrR5xlQtN3kc/view?usp=sharing
