- 博客(7)
- 收藏
- 关注
RSS订阅原创 C5eg1 Linear Regression
GoldTo build a system that can take a vector x⃗∈Rn\vec{x}\in\mathbb{R}^nx∈Rn as input and predict the value of a scalar y∈Ry\in\mathbb{R}y∈R as its output.DefinitionOutput:y^=w⃗Tx⃗\hat{y}=\vec{w...
2018-10-20 19:11:15
161
原创 C4eg1-Linear Least Squares
Suppose f(x)=12∥Ax−b∥22f(x)=\frac{1}{2}\|Ax-b\|_2^2f(x)=21∥Ax−b∥22gradient-based optimizationobtain the gradient: ∇xf(x)=AT(Ax−b)=ATAx−ATb\nabla_xf(x)=A^T(Ax-b)=A^TAx-A^Tb∇xf(x)=AT(Ax−b)=ATAx−A...
2018-10-20 19:08:20
133
原创 C4-Numerical Computation
Overflow and Underflowunderflowoccurs when numbers near zero are rounded to zeroOverflowoccurs when numbers with large magnitude are approximated as ∞\infty∞ or −∞-\infty−∞Softmax function...
2018-10-20 19:08:01
241
原创 C3-Probability and Information Theory
Possible sources of uncertaintyInherent stochasticity in the system being modeled.Incomplete observability.Incomplete modeling.ConceptsNote: In this note, x\text{x}x is random varaible and xxx ...
2018-10-20 19:06:56
205
原创 C2-Linear Algebra
conceptslinear algebra is a form of continuous rather than discrete mathematicsdetailed information --> Matrix Cookbook (Petersen and Pedersen, 2006)Scalars:–>(ponit)A scalar is just a sin...
2018-10-20 19:04:08
179
原创 C1-Introduction
Some ConceptsThe hierarcy of concepts allows the computer to learn complicated concepts by building them out of simpler ones. --> deep learningA computer can reason about statements in these fo...
2018-10-20 18:12:58
304
原创 C2eg1-Principal Components Analysis
Supposemmm points {x(1),⋯ ,x(m)}∈Rn\{\mathbf{x}^{(1)},\cdots,\mathbf{x}^{(m)}\} \in \mathbb{R}^n{x(1),⋯,x(m)}∈Rneach point x(i)∈Rn\mathbf{x}^{(i)}\in \mathbb{R}^nx(i)∈Rn corresponding...
2018-10-20 18:12:17
104
空空如也
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人