To serve as either a refresher or an introduction, Table A-1 summarizes the key mathematical terminology in the context in which it is used in this book.
Terminology | Description |
---|---|
x | A nonbold variable refers to a scalar. |
x | A variable in bold refers to a vector. |
y = f(x; Θ) | The result of function f on the vector input x, where f is dependent on the parameters Θ. In the context of this book, this represents the output of a DNN model for a particular input: f represents the DNN model algorithm, Θ represents its parameters determined during training, and x is the input to the model. |
C(f(x; Θ), y) | The result of function C given f(x; Θ) and the vector y. In the context of this book, this represents the cost (or loss) of the DNN model for a particular input with respect to the required output y. |
xi | The element i of vector x. |
The derivative of y with respect to x. | |
The partial derivative of y with respect to x, wh - ere x is one of the variables that affects y. | |
The nabla (upside-down Greek delta) symbol means “gradient.” refers to the vector of partial derivatives of the function f for the vector x. Put more simply, this means the effect that a very small change to the value of x has on the function f. | |
{1, 2, . . . L} | The set of numbers from 1 to L. |
ℝ | The set of real numbers. |
{x : P(x)} | The set of all values of x for which P(x) is true, where P(x) is a boolean statement. |
Returns x at which f(x) is minimized such that P(x) is true. | |
x ∈ ℝ | Indicates that x belongs to the set of real numbers. |
The Lp-norm of the vector x. Lp-norms are explained in “A Mathematical Approach to Measuring Perturbation”. | |
∑ x | The sum of all the possible values of x. |
The sum of all the possible values of x where . | |
ε | The Greek letter ε (epsilon) is used to indicate an extremely small (infinitesimal) quantity. |