When doing calibration using VQLS we need to express the matrix of the linear system as a sum of unitary matrices. This can be done brute force but that leads to (a small number) of very deep circuits.
In some cases (e.g. linear array) the matrix of the linear system reduces to an almost tridiagonal form that can be easily expressed as a sum of simple shallow circuits.
When doing calibration using VQLS we need to express the matrix of the linear system as a sum of unitary matrices. This can be done brute force but that leads to (a small number) of very deep circuits.
In some cases (e.g. linear array) the matrix of the linear system reduces to an almost tridiagonal form that can be easily expressed as a sum of simple shallow circuits.