Math & Machine Learning

It becomes important to understand some advanced math when you start to build robots that make use of some type of AI even if most of it relies on prewritten code like OpenCV , Tensorflow or OpenAI.

There are 4 main branches of mathematics that are in play with machine learning and they include linear algebra, calculus, statistics and probability and you could also throw in optimization.

We will break some of that down here in this video section.

But before we go there, you could watch a fascinating Youtube video about how important mathematics was to a US mathematics expert Nick Simons, known as the “king of quants” who learned the power of modelling, statistics and probability theory to make billions of dollars in the commodities and futures markets.

He started using his math skills as a code breaker and then branched out on his own into the money management business and discovered that models could be used to predict trends in the markets. He eventually became a billionaire using his skills with numbers.

But basically you must understand that AI is more popular today because we now have much more data available and AI is all about gathering and processing massive amounts of data.

They say that AI is so much more popular in 2021 than it was in the last 20 years because of these reasons.

  • much more data available often gathered by our search engine companies and social media
  • we have much more improved programming algorithms to make sense of the data
  • powerful processors are more accessible with GPU graphics processing engines

So with AI its all about gathering data and inputs and outputs from this data and putting it into rows and columns in tables and matrixes and using math to make sense of it all.

Eventually when you start getting into the mathematics of machine learning you will have a need to get some pre compiled batch of data….so at some point you may get the usual passenger survival data for something like the Titanic.

This area of analyzing vast amounts of data is hinted at in the movie Fight Club where the main star Ed Norton who plays the role of a disillusioned insurance analyst makes reference to the odds of survival and the types of injuries that are in the companies massive stored databases.

Facial recognition software is all about taking millions of photos and storing them and then seeing new faces and comparing some of the persons facial features with the database of existing faces. Comparisons are quickly made and questions are asked about the probability that this new person is a certain person that is in the existing database of maybe 100 million faces.

So lets talk about this handful of advanced mathematics areas just to give you a head start in them or at least to remind you what you might have slept thru way back in your college beer bash attending days.