Page 60 - Reclaim YOUR DIGITAL GOLD (without audio)

RECLAIM YOUR DIGITAL GOLD

learning. During this stage of the process, we will use the
data we have gathered to gradually improve our model’s
ability to determine whether a specific beverage is beer
or wine.

This is similar to how a person feels when they first obtain
their driver’s license. At first, they have no idea how to
use any of the pedals, knobs, or switches, let alone the
proper context in which to apply any of these controls.
A licensed driver, on the other hand, develops after a
significant amount of experience and the correction of
their errors. Furthermore, after a year of driving, they
have developed a higher level of expertise. Their driving
abilities have evolved as a result of driving and reacting
to real-world data, which has assisted them in honing
their skills.

This will be done on a smaller scale with our beverages.
To be more specific, y=m*x+b is the equation for a straight
line, where x is the input, m is the slope of the line, b is
the y-intercept, and y is the value of the line at x. The
values that we can use for “training” or “adjustment” are
m and b. There is no other way to influence the position
of the line because the only other variables are x, which
represents our input, and y, which represents our output.

There are many different m’s in machine learning because
there could be many different features. The sum of these
m values is commonly organized into a matrix, which
we will refer to as the “weights” matrix and denote with
the letter W. Similarly, we will group b and refer to the
resulting structure as the biases.

The training procedure starts with the selection of some
random values for W and b, and then attempts to predict

40