x
x
x |
标量 |
$x$ |
x
\boldsymbol{x}
x |
向量 |
$\boldsymbol{x}$ |
x
\mathbf{x}
x |
变量集 |
$\mathbf{x}$ |
A
\mathbf{A}
A |
矩阵 |
$\mathbf{A}$ |
I
\mathbf{I}
I |
单位阵 |
$\mathbf{I}$ |
X
\mathcal{X}
X |
状态空间 |
$\mathcal{X}$ |
D
\mathcal{D}
D |
概率分布 |
$\mathcal{D}$ |
D
D
D |
数据样本 |
$D$ |
H
\mathcal{H}
H |
假设空间 |
$\mathcal{H}$ |
H
H
H |
假设集合 |
$H$ |
L
\mathfrak{L}
L |
学习算法 |
$\mathfrak{L}$ |
(
⋅
,
⋅
,
⋅
)
(\cdot,\cdot,\cdot)
(⋅,⋅,⋅) |
行向量 |
$(\cdot,\cdot,\cdot)$ |
(
⋅
;
⋅
;
⋅
)
(\cdot;\cdot;\cdot)
(⋅;⋅;⋅) |
列向量 |
$(\cdot;\cdot;\cdot)$ |
(
⋅
)
T
(\cdot)^T
(⋅)T |
向量或矩阵转置 |
$(\cdot)^T$ |
{
⋯
}
\{\cdots\}
{⋯} |
集合 |
$\{\cdots\}$ |
∣
{
⋯
}
∣
\lvert \lbrace \cdots \rbrace \rvert
∣{⋯}∣ |
集合中元素个数 |
$\lvert \lbrace \cdots \rbrace \rvert$ |
∥
⋅
∥
p
\lVert \cdot \rVert _p
∥⋅∥p |
L
p
L_p
Lp 范数,
p
p
p缺省时为
L
2
L_2
L2范数 |
$\lVert \cdot \rVert _p$ |
P
(
⋅
)
,
P
(
⋅
∣
⋅
)
P(\cdot),\, P(\cdot \mid \cdot)
P(⋅),P(⋅∣⋅) |
概率质量函数, 条件概率质量函数 |
$P(\cdot),\, P(\cdot \mid \cdot)$ |
p
(
⋅
)
,
p
(
⋅
∣
⋅
)
p(\cdot),\,p(\cdot\mid\cdot)
p(⋅),p(⋅∣⋅) |
概率密度函数,条件概率密度函数 |
$p(\cdot),\,p(\cdot\mid\cdot)$ |
E
⋅
∼
D
[
f
(
⋅
)
]
\mathbb{E}_{\cdot \sim \mathcal{D}}\left [ f\! \left ( \cdot \right ) \right ]
E⋅∼D[f(⋅)] |
函数
f
(
⋅
)
f(\cdot)
f(⋅) 对
⋅
\cdot
⋅ 在分布
D
\mathcal{D}
D 下的数学期望; 意义明确时将省略
D
\mathcal{D}
D 和(或)
⋅
\cdot
⋅
|
$\mathbb{E}_{\cdot \sim \mathcal{D}}\left [ f\! \left ( \cdot \right ) \right ]$ |
sup
(
⋅
)
\sup(\cdot)
sup(⋅) |
上确界 |
$\sup(\cdot)$ |
I
(
⋅
)
\mathbb{I}(\cdot)
I(⋅) |
指示函数, 在
⋅
\cdot
⋅ 为真和假时分别取值为
1
,
0
1,0
1,0
|
$\mathbb{I}(\cdot)$ |
sign
(
⋅
)
\operatorname{sign}(\cdot)
sign(⋅) |
符号函数,在
⋅
<
0
,
=
0
,
>
0
\cdot\lt0,=0,\gt0
⋅<0,=0,>0时分别取值为
−
1
,
0
,
1
-1,0,1
−1,0,1
|
$\operatorname{sign}(\cdot)$ |
|
|
|