Inverse Matrices

If a matrix A is a transformation of space, its inverse A⁻¹ is the transformation that puts everything back. Do A then A⁻¹ and you return to where you started: A⁻¹A = A A⁻¹ = I, the identity.

This is the matrix version of dividing. There is no "÷ A"; instead you multiply by A⁻¹. It is what lets you solve Ax = b in one clean stroke — at least in principle.

Only square matrices can have inverses, and even then not all do. A matrix is invertible exactly when its determinant is non-zero. When the determinant is zero the matrix is singular: it collapsed space into a lower dimension, merging distinct inputs, and no transformation can un-merge them.

So "invertible", "det ≠ 0", "independent columns", "full rank" and "Ax = b always has a unique solution" are all the same statement — the good case from the previous lessons.

Keep an eye on the det readout. While it is non-zero the grid can be un-transformed back to the original — the matrix is invertible. Push it to 0 and the square collapses to a line: singular, no way back.

For a 2×2 matrix there is a formula worth memorising: swap the diagonal, negate the off-diagonal, and divide by the determinant.

Beyond 2×2 the practical method is Gauss–Jordan: write the matrix beside the identity as [A | I], row-reduce until the left side becomes I, and the right side has turned into A⁻¹.

One professional habit worth forming early: to solve Ax = b you rarely need the inverse at all. Elimination is faster and numerically safer than computing A⁻¹ and multiplying. Reach for the inverse when you truly need the matrix itself, not just one solution.