What math is needed for AI/ML R&D
When working on AI/ML research or needing to develop in this field, knowing the maths behind algorithms will always give you an advantage. There are many mathematical concepts to be understood when looking at AI/ML, so we will be breaking down the content in multiple sections.
If ML is an engine to predict the future, the data is the fuel. The data used in ML is often of high cardinality or dimensions. If the dataset is a table, each column represents a new dimension of the dataset, and each row is a sample of data. We can correlate each sample of the dataset as a vector in an N-dimensional coordinate plane, where N is the number of columns.
This makes our life easy, as in maths we have already defined various vector operations and calculations.
A linear map is a function applied to vector spaces such that it gives us back another vector
Example: consider x is a vector, a function f(x) is a function applied to a vector gives us back y, another vector.
f(x) = y
Why this helps us? Once we identify our dataset as a vector and know the Linear Maps transformation, we will get a set of tools that works for us to convert our dataset into another shape or form.