Probabilistic graphical models (PGMs) are a rich framework for encoding probability distributions over complex domains: joint (multivariate) distributions over large numbers of random variables that interact with each other. These representations sit at the intersection of statistics and computer science, relying on concepts from probability theory, graph algorithms, machine learning, and more. They are the basis for the state-of-the-art methods in a wide variety of applications, such as medical diagnosis, image understanding, speech recognition, natural language processing, and many, many more. They are also a foundational tool in formulating many machine learning problems. This course is the first in a sequence of three. It describes the two basic PGM representations: Bayesian Networks, which rely on a directed graph; and Markov networks, which use an undirected graph. The course discusses both the theoretical properties of these representations as well as their use in practice. The (highly recommended) honors track contains several hands-on assignments on how to represent some real-world problems. The course also presents some important extensions beyond the basic PGM representation, which allow more complex models to be encoded compactly.
Computing On Data
Loading...
From the course by Stanford University
Probabilistic Graphical Models 1: Representation
877 ratings
From the lesson
Bayesian Network (Directed Models)
In this module, we define the Bayesian network representation and its semantics. We also analyze the relationship between the graph structure and the independence properties of a distribution represented over that graph. Finally, we give some practical tips on how to model a real-world situation as a Bayesian network.
Meet the Instructors
Daphne KollerProfessor
School of Engineering
Now that you know how to load and save data and put your data in matrices and so
on, in this video I'd like to show you how to do computational operations on
data. And later on we'll, we'll be using the
source of computational operations to implement learning algorithms.
Let's get started. Here's my Octave window.
Let me just click the initialize some variables, to use for our examples.
Set A to be a two by two matrix, and set B to a two by two matrix, and let's set C
to a 2 by 2 matrix, like so. Now, let's say I want to multiply two of
my matrices. So, let's saw I want to compute a * c I
just type a * c. So it's a three by two matrix * a 2 by 2
matrix. This gives me this 3 by 2 matrix.
You can also do element wise operations. And do a.
* b. And what this will do is, it'll take each
element of a, and multiply it by the corresponding elements of b.
So that's a * b. That's a.
* b. So, for example, the first first element
gives. One times eleven which gives eleven.
The second element gives 2 * 12 which gives 24 and so on.
So does the element wise multiplication of two matrices and in general the p grid
tends to, is usually used to denote element wise operations and auto.
So here's a matrix A and Im going to do A.iii this gives me the mul- the element
wise squaring of A so you know, one square is one, two square is four and so
on. Let's set V to a vector, let's set V as
1, 2, 3 as a colon vector. You can also do one dot over V to do the
element Y's reciprocal of V. So this gives me 1 / 1, 1 / 2, and 1 / 3.
This works super matrices, so 1. / over a gives me that element y, your
inverse, of a. and once again the p-ratio gives us a
clue that this is elements wise operation.
You can also do things like log V, this is element wise logorithm of the of the,
the, V, E to the V, let's see of, base E exponentiation of v elements. So, this is
E, this is E squared, E cubed, because, this was V, so this, and,
I can also do absent V, to take the element wise absolute value of V, so
here, you know V was all positive to aps take minus one to minus three.
The element wise absolute value gives me back these non-negative values and
negative v gives me the minus of v. This is the same as negative one times v,
but usually you just write negative v instead of negative one times v.
And . What else can you do here is another neat
trick. So lets see, lets say I want to take V
and increment each others elements by one, well what we need to do is by
constructing a three x one. Vector this old ones and adding that to
V, so if I do that this increments v by from 1, 2, 3 to 2, 3, 4.
The way I did that was length of b is 3 so 1's length would be by one.
This is let 1's of 3 by 1 So, that's 1's three by one.
On the right, and what I did was V plus one is V by one, which is adding this
vector of all 1s to V and so this increments V by one and you, another
simpler way to do that, is actually type V plus one, right, so just V and V plus
one also means to add to one element wise to each of my elements of V.
Now. Let's talk about more operations.
So here's my matrix A. If you want to write A transpose the way
to do that is to write A prime. That's the, apostrophe symbol.
It's the left quote. So, in, in on your keyboard you probably
have a left quote, and a right quote. So this is, excuse.
This is actually a standard quotation mark.
Is, type A transpose this gives me the, you know, the transpose of my matrix A.
And of course, A transpose. If I transpose that again then I should
get back my matrix A. Some more useful functions let's say
lowercase a us 115 20.5, so it's a one by four matrix.
Let's say I say vow is max of A, this returns the maximum value of A which in
this case is fifteen and I can do vow end max A and this returns vow and end which
are going to be the maximum value of A which is fifteen as well as the index.
So is the element number two of A that was fifteen so ends is my index in this.
just as a warning that you do max A. Where A is a matrix, what this does is
this, this actually does the column wise maximum, but I'll say a little bit more
about this in a second. So using this example, the variable lower
case, a. If I do a less than three this does the
element y's operation, element y's comparison.
So the first element of a is less than three so this gives us one, second
element of a is not less than three so this device is zero cause it's false.
The third and fourth elements of r I meant less than three, 34 elements are
less than three so there's just one, one, so this does the elementwise comparison
of all four elements of the variable lowercase A to three and the returns true
or false depending on whether or not it's less than three.
Now, if I defined a less than three, this will tell me which are the elements of a.
The, the variable A are less than three, and in this case the first, third and
fourth elements are less than three. For my next example, let me set A to be
equal to magic to be the magic function, returns, let's say, help magic. The magic
function returns functions called magics returns these matrices call magic
squares, they have this, yeah mathematical property that's all of
their rows and columns and diagonals sum up to the same thing.
So you know, it's not actually useful for machine learning as far as I know but
I'm, I'm just using this as a convenient way.
You know, to generate a 3 by 3 matrix. And, and these magic squares anyway we
have the property that each row, each column and the diagonals all add up to
the same thing. So it's kind of a, mathematical
construct. I use magic, I use this magic function only when I'm doing demos or
when I'm teaching Octave, like this and I don't actually use it for any, you know
useful machine only application. But let's see if I type RC equals find A
greater than or equals seven, this finds all the elements of A, that are greater
than or equal to seven. And so, RC sends row and columns.
So the 1-1 element is greater than seven. The 3-2 element is greater than seven and
the 2-3 element is greater than seven. So let's see.
The 2-3 element, for example, is a 2-3 is seven is this element.
Out here, and that is, indeed, greater than or equal to seven.
By the way, I actually don't even memorize, myself, what these fine
functions do. And what all of these things do myself.
And whenever I use the fine function, sometimes I forget myself exactly what it
does. And you know, I to, like, help find, to
look up the document. Okay just two more things and I'll
quickly show you, one is the sum function.
So here's my A I'm going to type sum A which adds up all the elements of A and
if I want to multiply them together I type prod, A prod sense of product and
this returns the product of these four elements of A.
four A rounds down these elements of A so 0.5 is rounded down to zero and seal or
sealing A gets rounded up to the nearest integers so 0.5 gets rounded up to 1.
You can also let's see. Let me type rand(3), this generates a 3
by 3 matrix. If I type max rand(3), rand(3) what this
does is it takes the element wise maximum of a two random 3 by 3 matrices.
So, you notice, all of these numbers tend to be a bit on the large side.
Because each of these is actually the max of a randomly of, of element Y's max of
two randomly generated matrices. This is, this was my magic number, this
was my magic square three by three A. Let's say I type max A
and then this will be eight. Open close square brackets, comma one.
What this does is this takes the column-wise maximum.
So, the max of the first column is eight, the max of the second column is nine, and
the max in the third column is seven. This one means to take the max along the
first dimension of A. In contrast if I were to take max A, this
funny notation too, then this takes this per row maximum, so the max of the first
row is eight, max of second row is seven, max of the third row is nine and so this
allows you to take maxes, either, you know, per row, or per column.
And if you want to and remember, it defaults to a column like wise element.
So you won't find. If you want to find the maximum element
in the entire matrix A. You can type max of max of A.
Like so, which is 9 or you can turn A into vector and type max of A: like so.
And this treats this as a vector. And takes the max of, max elements of
that vector. Finally,
let's set A to be a 9 by 9 magic square, so remember the magic square has this
property that every column and every row sums the same thing and also the
diagonals. So here's a nine by nine may magic
square, so let me do sum A1 so this does a per column sum so I'm going to take
each column of A and ad them up and this, you know let's just verify that indeed
for a nine by nine magic square every column adds up to 369, adds up to the
same thing. Now let's do the row-wide sum so the sum
A, comma 2 and this sums up each row of A, and indeed each row of A also sums up
to 369. Now, let's sum the diagonal elements of a
and make sure that they, that, that also sums to the same thing.
So, what I'm going to do is construct a 9 by 9 identity matrix.
That's I9 and I'm going to take a and construct, you know multiply A,
element-wise. So here's my matrix A.
I'm going to do a * I9. And what this will do is take the
element-wise product of these two matrices.
And so this should wipe out everything in a except for the diagonal entries.
And now I'm going to do sum, sum of a of that.
And this gives me the sum of this, these diagonal elements.
And indeed, it is 369 you can sum up the other diagonal as well.
So the sort of top left to bottom right, you can sum up the opposite diagonal from
bottom left to top right. the sum, the, the commands for this is
somewhat more cryptic. You don't need to know this.
I'm just showing you this in case any of you are curious.
But let's sese flip U, flip UD stands for flip up down.
But if you do that, that turns out to sum up the elements and the opposites of the
other diagonal that also sums up to 369. Here let me show you.
Whereas I9 is this matrix. Flip up down of I9 you know takes the
identity matrix and flips it vertically. So you end up with, excuse me flip UD end
up with 1's on this opposite diagonal as well.
Just one last command and then that's it, and then that'll be it for this video.
Let's set A to be the magic, three by three magic square game.
If you want to invert a matrix, you type P inv A.
this is typically called the pseudo inverse, but it doesn't matter.
Just think of it as basically the inverse of A, and that's the inverse of A.
And so I can set, you know, temp equals P inv of A and [INAUDIBLE] temp times A.
This is indeed the identity matrix with essentially ones on the diagonals and
zero. It was on the off diagonals up to a
numerical round off. So that's it for how to do different
computational operations on the data matrices.
and after running a learning algorithm, often one of the most useful things is to
be able to look at your results. Or to plot or visualize your results.
And in the next video I'm going to very quickly show you how again with one or
two lines of code, using Octave. You can quickly visualize your data or plot your
data. And use that to better understand you
know, what your learning algorithms are doing.
Explore our Catalog
Join for free and get personalized recommendations, updates and offers.
Coursera provides universal access to the world’s best education,
partnering with top universities and organizations to offer courses online.
© 2018 Coursera Inc. All rights reserved.
