ML - 7 - Linear Algebra
Tariq’s introduction to machine learning gave me a solid foundation of understanding. As I have been reading through Michael’s machine learning book, it has become readily apparent that a dive into linear algebra will benefit me.
Funny to think that someone with an engineering degree never really enjoyed math during school. Linear Algebra was an elective, so I skipped taking it. There were other justifications behind it. Namely that it was my last semester and I had other classes that I had to pass in order to graduate. I didn’t want to add a heavy workload course that was not necessary to graduate.
A few years after college I was introduced to Khan Academy. Khan’s explanations of certain math subjects resonated with me. On top of that I read a couple books by Steven Strogatz that made me see mathematics in a different light.
I’m not at the ‘proof’ level with math, but I definitely get more enjoyment out of it.
That being said, I am going to dig into Linear Algebra on Khan Academy’s website. I think this will help me follow along with Michael’s book. As well as the explanations of exercises provide by Tom Fahey on GitHub.
In the past I would have seen this type of detour(for lack of a better word) as a negative. Now I follow my interest. If I am enjoying the subject and excited about a deeper level of understanding, then time is not an issue. I have no deadline.
Right off the bat I see how LA could help better understand ML.
Linear algebra allows us to apply our understanding and ability to visualize 1,2, and 3 dimensions to higher order dimensions.
Real Coordinate Space:
a vector includes all of the real number possibilities for n dimensions
There are probably programs that can visualize higher order real coordinate spaces.
I wonder what, if anything, changes from a calculation standpoint when the problems move higher than 1, 2, or 3 dimensions. My expectation is that what I learn for lower order will apply to higher order. Even without my ability to visualize I will be able to understand higher order linear algebra. Not being able to visualize would have hurt me in the past because I like to work things visually in my mind.
Tuple - Ordered list
I have seen this word related to python, but I am not sure that I have seen this specific definition. Or rather, did not recognize it when I saw it.
Sal goes through adding vectors algebraically and graphically. Nothing too mind blowing here but necessary.
Algebraically: Add rows to get result.
Graphically: Drawing vectors nose to tail. Either vector can be drawn first.
The next video I am going to check out is: