~/writing/math-foundations-for-reasoning-models $ ls
Probability for Learning
Random Variable
A random variable is a value-producing variable whose result is uncertain before observation.
~/writing/math-foundations-for-reasoning-models $ cat topic.md
Random Variable
Definition
A random variable is a variable whose value is uncertain before it is observed.
It represents a possible outcome using a value.
Example:
X = result of a coin flip
Possible values:
heads
tails
Random variable = uncertain outcome represented as a value.
Why It Is Called A Variable
It is called a variable because it can take different values.
Example:
X = number shown after rolling a die
Possible values:
1, 2, 3, 4, 5, 6
Before the roll, we do not know which value X will take.
After the roll, the value is observed.
A random variable names the uncertain quantity we care about.
Random Does Not Mean Patternless
Random does not mean there is no structure.
For a fair die:
each face has probability 1/6
The exact result is uncertain.
But the possible values and their probabilities can still be structured.
Probability studies structured uncertainty, not pure chaos.
Random Variable vs Value
A random variable is the uncertain quantity.
A value is one realized outcome.
Example:
X = die roll
X = 4 after observing the roll
Before observation:
X could be 1, 2, 3, 4, 5, or 6
After observation:
X has one value
Random variable before observation; value after observation.
Discrete Random Variable
A discrete random variable has separate possible values.
Example:
number of emails received today
Possible values:
0, 1, 2, 3, ...
The values are countable.
Discrete random variables take separate values.
Continuous Random Variable
A continuous random variable can take values along a continuum.
Example:
temperature
time
height
Possible values are not just separate points.
There can be values between values.
Continuous random variables take values over a continuous range.
Why Random Variables Matter For Learning
Learning systems often deal with uncertainty.
Example:
X = input features
Y = label
The model may not know the correct label with certainty.
Instead, it may estimate probabilities.
Random variables let us talk about uncertain inputs, labels, and predictions.
Final Summary
- A random variable represents an uncertain outcome as a value.
- It can take different possible values before observation.
- Random does not mean patternless.
- A random variable is not the same as one realized value.
- Random variables can be discrete or continuous.
- Learning uses random variables to reason about uncertain data and labels.
~/writing/math-foundations-for-reasoning-models $ open
Continue the series.
Probability Distribution