[MUSIC] But before we delve into the course
itself, let's just talk about what we're assuming you have as background in order
to successfully complete this course. So in terms of math background,
we're assuming that you know very basic calculus like how to take
the derivative of a single variable. Function. But even that we're gonna walk
through fairly slowly, but it's good if you remember it. We're also gonna assume that you know
how to do very basic linear algebra manipulations like how to
multiply to vectors, to matrices, and also of course what the notion is of
a matrix and a vector, but again, we're gonna step through these ideas fairly
slowly at the beginning of the course. In terms of programming experience,
we've tried to make this course as open as possible to people having
preferences in different languages. We're gonna, encourage the use of Python. But this is not actually required. So all of our programming assignments,
we're gonna provide you with some starter code and
that starter code will be in Python. So of course it's helpful if you
are familiar with Python, but again you're welcome to use
any language you would like. We're focused on teaching you
the concepts of machine learning rather than any specific
implementation details. But, I should mention that if you
did programming assignments for the first course, you are set for
what will be required in this course. So in the first course
of this specialization, we relied on pre-implemented algorithms
like GraphLab Create, but in this course, we're actually going to teach you all
the all the algorithmic details and how to implement these algorithms so
that you can code these up yourself. So in particular, in this course,
we're suggesting that you use SFrames, which from the first course you remember
is an open source library that allows for really scalable and efficient data
manipulation, but you're welcome to use any data structure that you would like or
library such as pandas. And for our assignments. The structure of the assignments are going
to start with exploring high level concepts in the first part and then delving into that implementation
details of each of the algorithms. So for the first part, we're going to
encourage the use of pre-implemented algorithms, so that you can test these
high level concepts without getting bogged down in potential bugs in
your actual implementation. And then in the second part,
once you've explored those concepts, then we're gonna have you actually
implement all the algorithms from scratch without relying on these
pre-implemented methods. If you're using Python,
you will be using the Numpy library for manipulating matrices and vectors. Okay, so
the net result is that in this course, you're really gonna get your hands
dirty in implementing all the methods. Related to regression that you saw in
the first foundational course this specialization as well as a bunch
of new methods that we're gonna explore in this course. So finally, let's discuss what computing resources
you guys need and there are two options. Either, you have your own computer and it
can actually be a fairly basic computer. But if you are gonna use SFrames,
you are gonna need a 64-bit machine. You'll also need to have access to
the internet, of course, to watch these wonderful videos, as well as to
implement your programming assignments. And you'll need the ability to install and
run Python. And any libraries we're using
associated with that and to store a few gigabytes of data, but
the alternative, especially if you don't have a 64-bit machine, is we're
providing a set of machines in the cloud that are pre-configured with
everything you'll need for this course. So all you'll need is internet access
to access those machines in the cloud. So now that you guys are psyched
about this course, let's get started. [MUSIC]