Linear-Algebra-Done-Right · Unit 17 · Video 2 · Interactive Practice
| Formula | Name | What it says |
|---|---|---|
| Restriction theorem | Injective and onto , hence invertible | |
| Pseudoinverse | Project onto , then invert the restriction | |
| invertible | Property (1) | Nothing changes when an inverse exists |
| , | Properties (2) and (3) | Both products are orthogonal projections |
Key Insight: sends to , and sends each to the one vector of that maps to .
For , projects onto , then pulls it back into .
💡 The solutions of form the line ; the next video proves that is the one of smallest norm.
Sending to and back, keeps the piece of and deletes the piece.
and are orthogonal projections, and each is an identity exactly when is surjective or injective.
Problem 1 · Compute a Pseudoinverse Value
Given: , — find .
, so .
, so .
Find with :
Hence . The other solutions of , such as , differ from it by a vector of .
Problem 2 · Injective but Not Surjective
Given: is finite-dimensional and is injective but not surjective — which statement is true?
injective means , so and property (3) gives
not surjective means , so property (2) gives , which sends every nonzero vector of to and is therefore not .
So is a left inverse of but not a right inverse.
Problem 3 · A Map from F³ to F²
Given: , — find and .
What is ?
What is ?
is surjective (it sends and to the standard basis), so and .
: and , so and .
Part 1. Solve , , . From the first two, and ; then gives :
Part 2. By property (3), . With and :
Check: and , and .
Problem 4 · The Pseudoinverse of a Projection
Given: is a subspace of a finite-dimensional inner product space with , and — find .
For : and , so .
The restriction sends each to , so it is the identity on and its inverse is the identity on too. By the definition,
Hence . Consistently with properties (2) and (3), and .
Solved: 0 / 4