I have a matrix of the form,

$$ \begin{bmatrix} x_1^\top A x_1, \ldots, x_1^\top A x_n \\ \cdots, \cdots, \cdots, \\ x_m^\top A x_1, \ldots, x_m^\top A x_n \end{bmatrix} $$

where $A$ is some matrix, and $x_i$'s are column vectors with the appropriate dimensions for the matrix $A$. Note that each element of the above matrix is a scalar in the quadratic form $x_i^\top A x_j$.

Question: Is there a nice compact notation to express the above?

I thought about using Kronecker product; say $B \otimes C $ has elements that look like $b_{ij} C$, which is close to what I'm looking for, except that I have a scalar $b_{ij}$ multiplying the matrix $C$ instead of a vector-times-matrix product.



You can describe this as

$$ \begin{pmatrix} x_1^T \\ \vdots \\ x_m^T \end{pmatrix} A \begin{pmatrix} x_1 & \dots & x_n \end{pmatrix} = \begin{pmatrix} x_1^T \\ \vdots \\ x_m^T \end{pmatrix} \begin{pmatrix} Ax_1 & \dots & Ax_n \end{pmatrix}. $$

Let's say $A \in M_{l \times l}(\mathbb{F})$. The matrix $\begin{pmatrix} x_1 & \dots & x_n \end{pmatrix}$ is the $l \times n$ matrix whose $i$-th column is $x_i$.

$$ \begin{pmatrix} x_1^T \\ \vdots \\ x_m^T \end{pmatrix} $$

is a $m \times l$ matrix whose $i$-th row is $x_i^T$. By block decomposition, the product is your required matrix.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.