Review

Review

Chapter 1: Linear Systems

Linear Systems (ch1Lecture1b, ch1Lecture2)

• Applications of linear systems
• Putting linear systems in matrix form
• *Gauss-Jordan to get to row echelon form
• *Solving linear systems with augmented matrices
• Free vs bound variables
• Ill-conditioned systems & rounding errors

Getting to row echelon form: \[ \left[\begin{array}{rrr} 1 & 1 & 5 \\ 0 & -3 & -9 \end{array}\right] \overrightarrow{E_{2}(-1 / 3)}\left[\begin{array}{lll} 1 & 1 & 5 \\ 0 & 1 & 3 \end{array}\right] \overrightarrow{E_{12}(-1)}\left[\begin{array}{lll} 1 & 0 & 2 \\ 0 & 1 & 3 \end{array}\right] . \]

Augmented matrix to solve linear system:

\[ \begin{aligned} z & =2 \\ x+y+z & =2 \\ 2 x+2 y+4 z & =8 \end{aligned} \]

. . .

Augmented matrix:

\[ \left[\begin{array}{llll} 0 & 0 & 1 & 2 \\ 1 & 1 & 1 & 2 \\ 2 & 2 & 4 & 8 \end{array}\right] \stackrel{E_{12}}{\longrightarrow}\left[\begin{array}{llll} (1 & 1 & 1 & 2 \\ 0 & 0 & 1 & 2 \\ 2 & 2 & 4 & 8 \end{array}\right] \xrightarrow[E_{31}(-2)]{\longrightarrow}\left[\begin{array}{rlll} 1 & 1 & 1 & 2 \\ 0 & 0 & 2 & 4 \\ 0 & 0 & 1 & 2 \end{array}\right] \]

We keep on going… \[ \begin{aligned} & {\left[\begin{array}{rrrr} (1) & 1 & 1 & 2 \\ 0 & 0 & 2 & 4 \\ 0 & 0 & 1 & 2 \end{array}\right] \xrightarrow[E_{2}(1 / 2)]{\longrightarrow}\left[\begin{array}{rrrr} 1 & 1 & 1 & 2 \\ 0 & 0 & 1 & 2 \\ 0 & 0 & 1 & 2 \end{array}\right]} \\ & \overrightarrow{E_{32}(-1)}\left[\begin{array}{rrrr} (1) & 1 & 1 & 2 \\ 0 & 0 & 1 & 2 \\ 0 & 0 & 0 & 0 \end{array}\right] \xrightarrow[E_{12}(-1)]{\longrightarrow}\left[\begin{array}{rrrr} (1) & 1 & 0 & 0 \\ 0 & 0 & 1 & 2 \\ 0 & 0 & 0 & 0 \end{array}\right] . \end{aligned} \]

There’s still no information on \(y\).

\[ \begin{aligned} & x=-y \\ & z=2 \\ & y \text { is free. } \end{aligned} \]

Chapter 2: Matrices

Matrix multiplication (ch2Lecture1)

• Matrix-vector multiplication as a linear combination of columns
• Matrix multiplication as an operation
• *Scaling, rotation, shear

Scaling and rotation

To rotate a vector by \(\theta\):

\[ =\left[\begin{array}{rr} \cos \theta &-\sin \theta \\ \sin \theta & \cos \theta \end{array}\right]\left[\begin{array}{l} r \cos \phi \\ r \sin \phi \end{array}\right]=\left[\begin{array}{rr} \cos \theta&-\sin \theta \\ \sin \theta & \cos \theta \end{array}\right]\left[\begin{array}{l} x \\ y \end{array}\right] \]

. . .

Scaling:

\[ A = \left[\begin{array}{ll} z_1 & 0 \\ 0 & z_2 \end{array}\right] \]

Shearing: adding a constant shear factor times one coordinate to another coordinate of the point. \[ A = \left[\begin{array}{ll} 1 & s_2 \\ s_1 & 1 \end{array}\right] \]

Graphs and Directed Graphs (ch2Lecture2)

• *Adjacency and incidence matrices
• Degree of a vertex
• *Paths and cycles
• PageRank

Adjacency and incidence matrices

Adjacency matrix: A square matrix whose \((i, j)\) th entry is the number of edges going from vertex \(i\) to vertex \(j\)

Incidence matrix: A matrix whose rows correspond to vertices and columns correspond to edges. The \((i, j)\) th entry is 1 if vertex \(i\) is the tail of edge \(j\), -1 if it is the head, and 0 otherwise.

Paths and cycles

• Number of paths of length \(n\) from vertex \(i\) to vertex \(j\) is the \((i, j)\) th entry of the matrix \(A^{n}\).
• Vertex power is the sum of the entries in the \(ith\) row of \(A+A^{2}\).
• In an the incidence matrix of a digraph which is a cycle, every row must sum to zero.

Discrete Dynamical Systems (Ch2Lecture2)

• Transition matrices
• *Markov chains

Markov Chains

A distribution vector is a vector whose entries are nonnegative and sum to 1.

A stochastic matrix is a square matrix whose columns are distribution vectors.

A Markov chain is a discrete dynamical system whose initial state \(\mathbf{x}^{(0)}\) is a distribution vector and whose transition matrix \(A\) is stochastic, i.e., each column of \(A\) is a distribution vector.

Difference Equations (Ch2Lecture3)

• *Difference equations in matrix form
• *Examples

Difference Equation in Matrix Form

From HW2: put ths difference equation in matrix form:

\[ y_{k+2}-y_{k+1}-y_{k}=0 \]

. . .

Steps: 1. Make two equations. Solve for \(y_{k+2}\): \(y_{k+2}=y_{k+1}+y_{k}\). Also of course \(y_{k+1}=y_{k+1}\). 2. Define the vector \(\mathbf{y}^{(k)}=\left[\begin{array}{l} y_{k} \\ y_{k+1} \end{array}\right]\) 3. Put the two equations in matrix form: . . .

\(\left[\begin{array}{l}y_{k+1} \\ y_{k+2}\end{array}\right]=\left[\begin{array}{ll}0 & 1 \\ 1 & 1\end{array}\right]\left[\begin{array}{c}y_{k} \\ y_{k+1}\end{array}\right]\)

Examples of difference equations

Reaction-diffusion:

\[ a_{x,t+1} = a_{x,t} + dt\left( \frac{D_{a}}{dx^{2}}(a_{x+1,t} + a_{x-1,t} - 2a_{x,t}) \right) \]

Heat in a rod: \[ -y_{i-1}+2 y_{i}-y_{i+1}=\frac{h^{2}}{K} f\left(x_{i}\right) \]

MCMC (Ch2 lecture 4)

• MCMC
  • Simulate a distribution using a Markov chain
  • Sample from the simulated distribution
• Restricted Boltzmann Machines

Inverses and determinants (Ch2 lecture 4 & 5)

• Inverse of a matrix
• *Determinants
• Singularity
• LU factorization

Determinants

The determinant of a square \(n \times n\) matrix \(A=\left[a_{i j}\right]\), \(\operatorname{det} A\), is defined recursively:

If \(n=1\) then \(\operatorname{det} A=a_{11}\);

otherwise,

• suppose we have determinents for all square matrices of size less than \(n\)
• Define \(M_{i j}(A)\) as the determinant of the \((n-1) \times(n-1)\) matrix obtained from \(A\) by deleting the \(i\) th row and \(j\) th column of \(A\)

then

\[ \begin{aligned} \operatorname{det} A & =\sum_{k=1}^{n} a_{k 1}(-1)^{k+1} M_{k 1}(A) \\ & =a_{11} M_{11}(A)-a_{21} M_{21}(A)+\cdots+(-1)^{n+1} a_{n 1} M_{n 1}(A) \end{aligned} \]

Laws of Determinants

• D1: If \(A\) is an upper triangular matrix, then the determinant of \(A\) is the product of all the diagonal elements of \(A\).
• D2: If \(B\) is obtained from \(A\) by multiplying one row of \(A\) by the scalar \(c\), then \(\operatorname{det} B=c \cdot \operatorname{det} A\).
• D3: If \(B\) is obtained from \(A\) by interchanging two rows of \(A\), then \(\operatorname{det} B=\) \(-\operatorname{det} A\).
• D4: If \(B\) is obtained from \(A\) by adding a multiple of one row of \(A\) to another row of \(A\), then \(\operatorname{det} B=\operatorname{det} A\).
• D5: The matrix \(A\) is invertible if and only if \(\operatorname{det} A \neq 0\).
• D6: Given matrices \(A, B\) of the same size, \(\operatorname{det} A B=\operatorname{det} A \operatorname{det} B \text {. }\)
• D7: For all square matrices \(A\), \(\operatorname{det} A^{T}=\operatorname{det} A\)

Chapter 3: Vector Spaces

Spaces of matrices (ch3 lecture 1)

• Basis
• Fundamental subspaces:
  • Column space
  • Null space
  • Row space
• Rank
• Consistency

Column and Row Spaces

The column space of the \(m \times n\) matrix \(A\) is the subspace \(\mathcal{C}(A)\) of \(\mathbb{R}^{m}\) spanned by the columns of \(A\).

The row space of the \(m \times n\) matrix \(A\) is the subspace \(\mathcal{R}(A)\) of \(\mathbb{R}^{n}\) spanned by the transposes of the rows of \(A\)

. . .

A basis for the column space of \(A\) is the set of pivot columns of \(A\). (Find these by row reducing \(A\) and choosing the columns with leading 1s)

Null Space

The null space of the \(m \times n\) matrix \(A\) is the subset \(\mathcal{N}(A)\) of \(\mathbb{R}^{n}\)

\[ \mathcal{N}(A)=\left\{\mathbf{x} \in \mathbb{R}^{n} \mid A \mathbf{x}=\mathbf{0}\right\} \]

. . .

\(\mathcal{N}(A)\) is just the solution set to \(A \mathbf{x}=\mathbf{0}\)

. . .

For example, if \(A\) is invertible, \(A \mathbf{x}=\mathbf{0}\) has only the trivial solution \(\mathbf{x}=\mathbf{0}\)

. . .

so \(\mathcal{N}(A)\) is just \(\left\{\mathbf{0}\right\}\).

. . .

\(A\) is invertible exactly if \(\mathcal{N}(A)=\{\mathbf{0}\}\)

Finding a basis for the null space

Given an \(m \times n\) matrix \(A\).

1. Compute the reduced row echelon form \(R\) of \(A\).

2. Use \(R\) to find the general solution to the homogeneous system \(A \mathbf{x}=0\).

. . .

3. Write the general solution \(\mathbf{x}=\left(x_{1}, x_{2}, \ldots, x_{n}\right)\) to the homogeneous system in the form

. . .

\[ \mathbf{x}=x_{i_{1}} \mathbf{w}_{1}+x_{i_{2}} \mathbf{w}_{2}+\cdots+x_{i_{n-r}} \mathbf{w}_{n-r} \]

where \(x_{i_{1}}, x_{i_{2}}, \ldots, x_{i_{n-r}}\) are the free variables.

. . .

4. List the vectors \(\mathbf{w}_{1}, \mathbf{w}_{2}, \ldots, \mathbf{w}_{n-r}\). These form a basis of \(\mathcal{N}(A)\).

Using the Null Space

• The general solution to \(A \mathbf{x}=\mathbf{b}\) is \(\mathbf{x}=\mathbf{x}_{p}+\mathbf{x}_{h}\) where \(\mathbf{x}_{p}\) is a particular solution and \(\mathbf{x}_{h}\) is in the null space of \(A\).
• The null space of \(A\) is orthogonal to the row space of \(A\) (or the column space of \(A^{T}\). The dot product of any vector in the null space of \(A\) with any vector in the row space of \(A\) is 0.)

Chapter 4: Geometrical Aspects of Standard Spaces

Orthogonality (ch4 lecture 1 and 2)

• Geometrical intuitions
• *Least Squares & Normal equations
• Finding orthogonal bases (Gram-Schmidt)

Least Squares and Normal Equations

To find the least squares solution to \(A \mathbf{x}=\mathbf{b}\), we minimize the squared error \(\left\|A \mathbf{x}-\mathbf{b}\right\|^{2}\) by solving the Normal Equations for \(\mathbf{x}\): \[ \begin{aligned} \mathbf{A}^{T} \mathbf{A} \mathbf{x} &=\mathbf{A}^{T} \mathbf{b} \end{aligned} \]

QR Factorization

If \(A\) is an \(m \times n\) full-column-rank matrix, then \(A=Q R\), where the columns of the \(m \times n\) matrix \(Q\) are orthonormal vectors and the \(n \times n\) matrix \(R\) is upper triangular with nonzero diagonal entries.

1. Start with the columns of A, \(A=\left[\mathbf{w}_{1}, \mathbf{w}_{2}, \mathbf{w}_{3}\right]\). (For now assume they are linearly independent.)
2. Do Gram-Schmidt on the columns of \(A\) to get orthonormal vectors \(\mathbf{q}_{1}, \mathbf{q}_{2}, \mathbf{q}_{3}\).

. . .

\[ A=\left[\mathbf{w}_{1}, \mathbf{w}_{2}, \mathbf{w}_{3}\right]=\left[\mathbf{q}_{1}, \mathbf{q}_{2}, \mathbf{q}_{3}\right]\left[\begin{array}{ccc} 1 & \frac{\mathbf{q}_{1} \cdot \mathbf{w}_{2}}{\mathbf{q}_{1} \cdot \mathbf{q}_{1}} & \frac{\mathbf{q}_{1} \cdot \mathbf{w}_{3}}{\mathbf{q}_{1} \cdot \mathbf{q}_{1}} \\ 0 & 1 & \frac{\mathbf{q}_{2} \cdot \mathbf{w}_{3}}{\mathbf{q}_{2} \cdot \mathbf{q}_{2}} \\ 0 & 0 & 1 \end{array}\right] \]

Chapter 5: Eigenvalues and Eigenvectors

Eigenvalues and Eigenvectors (ch5 lecture 1)

• Definition
• *How to find them
• Similarity and Diagonalization
• Applications to dynamical systems
• Spectral radius

Finding Eigenvalues and Eigenvectors

If \(A\) is a square \(n \times n\) matrix, the equation \(\operatorname{det}(\lambda I-A)=0\) is called the characteristic equation of \(A\)

The eigenvalues of \(A\) are the roots of the characteristic equation.

For each scalar \(\lambda\) in (1), use the null space algorithm to find a basis of the eigenspace \(\mathcal{N}(\lambda I-A)\).

Symmetric matrices (ch5 lecture 2)

• Properties of symmetric matrices
• Quadratic forms

SVD (ch5 lecture 3, 4)

• Definition
• *Psuedoinverse
• Applications to least squares
• Image compression

Pseudoinverse

The pseudoinverse of \(A\) is \(A^{+}=V S^+ U^{T}\)

\(S^{+}\) is the matrix with the reciprocals of the non-zero singular values on the diagonal, and zeros elsewhere.

. . .

Can find least squares solutions:

\[ \begin{aligned} A x & =b \\ x & \approx A^{+} b \end{aligned} \]

PCA (ch5 lecture 5)

• Definition
• Applications to data analysis

Chapter 6: Fourier Transform

Fourier Transform (ChNone, Ch6 lecture 1)