# Notation

This note contains the notations/terms used in different contexts so as to be consistent.

## General

### Note titles:

- In Title Case, see here.
- Should contain plural form unless referring to specific instance, e.g.
`Multi-Armed Bandits (MABs)`

instead of`Multi-Armed Bandit (MAB)`

, and`Kalman Filter`

instead of`Kalman Filters`

. - For things with multiple names, look into roam aliases.

### Lists:

- For things like properties, state the name followed by the mathematical description, then any conditions it should satisfy. E.g. =Associativity: \((a+b)+c=a+(b+c)\) assuming finite \(a,b,c\).

### Content:

- For things with multiple names, do the following, e.g.
`Contextual bandits (also called associative search)...`

where the first name is the one you see used most commonly/used by leading researchers. - Fullstop after each sentence, even in lists.
- When listing names of mathematical properties, put their names in parentheses after stating it in math e.g. \(\x \in \mathcal{X},x\geq 0\) (Non-negativity).

### Math/LaTeX:

- Argmax/min written using math text e.g. \(\text{argmax}\).

## Reinforcement Learning

- Uppercase for random variables, lowercase for outcomes.
- (Experimental) Hat notation for estimates of e.g. Q functions.
- State-value instead of State value.

## Bayesian statistics

- Unnormalized densities have a tilde, e.g. \(\tilde{\pi}\)
- Normalizing constant of density \(p\) is denoted \(Z_p\).