1
$\begingroup$

I am currently reading Deep Learning book, and I want to get better understanding of probability theory. In chapter 3.3.1 of Deep Learning book it states that:

Often we associate each random variable with a different probability mass function and the reader must infer which PMF to use based on the identity the random variable, rather than on the name of the function;P(x) is usually not the same as P (y).

And not many paragraphs latter, it saids the following:

Probability mass functions can act on many variables at the same time. Such a probability distribution over many variables is known as a joint probability distribution.P(x=x, y=y) denotes the probability that x=x and y=y simultaneously. We may also write P (x, y) for brevity

I am having hard time grasping these two paragraphs. What do they mean when they say that P(x) is USUALLY not the same as P(y). As I understood, random variable is basically a random phenomenon from a real world that we wish to model. And each random phenomenon has its own Probability Mass function. Does this mean that the first paragraph indicates that random variable y represents a different phenomenon, and in the second paragraph random variable y represents the same type of phenomenon as x and that is why we use the same probability mass function?

Thanks in advance!

$\endgroup$

1 Answer 1

1
$\begingroup$

I believe they mean the following: "We will use $P$ to denote the concept of a probability mass function (pmf). $P(x)$ is the pmf of random variable $x$, while $P(y)$ is the pmf of random variable $y$. The two variables are usually the results of different random experiments, hence we use two separate letters, $x$ and $y$. As such, they will probably have different pmfs. But somethimes, it might happen that they have the same."

I think the essential source of confusion is that $P$ is not used to denote any single pmf, but is the name used to refer to a pmf. As a parallel, height_of(Stefan) and height_of(Mark) may be different, or they may be the same. height_of is just the name of the thing we are measuring, not the thing itself.

P(x,y) is thus the pmf of the joint variable (x,y). P((x, y)) would be a more unambiguous notation, but I guess is omitted for brevity. If you need more clarification on joint variables let me know. I hope this helps.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.