OE-4.1 Math Foundations for Quantum Computing

What this is and why it exists

Quantum computing is linear algebra with a physical interpretation attached. This unit is the algebra, with the physics deliberately withheld.

That withholding is a kindness. Learn the mathematics as mathematics and the physics arrives later as meaning rather than as mystery.

One pair gets confused constantly and every later chapter depends on telling them apart: the inner product and the outer product. One produces a number, the other produces an operator.

The vocabulary

• Vector space — a set of vectors you can add and scale.
• Linear independence — no vector in the set is a combination of the others.
• Basis — a smallest set from which every vector can be built.
• Orthogonal — two vectors whose inner product is zero.
• Inner product — two vectors in, one number out. A measure of overlap.
• Outer product — two vectors in, an operator out.
• Hilbert space — a vector space with an inner product, complete in the sense analysis requires.
• Unitary operator — one that preserves lengths and angles.
• Projection — an operator that drops a vector onto a subspace.
• Eigenvector — a vector an operator only stretches, never turns.
• Eigenvalue — the factor by which it is stretched.

The mental model

A basis is a set of directions from which everything else can be built, and the number of them is the dimension. That is the whole idea, and once you have it, "a state is a vector" stops being intimidating.

Now the pair that matters. The inner product takes two vectors and returns a single number, which measures how much they overlap. Two orthogonal vectors have inner product zero: no overlap at all. The outer product takes the same two vectors and returns an *operator* — a thing that acts on vectors. Same two inputs, completely different kind of output. If you remember only one sentence from this topic, remember that one, because later chapters build operators out of outer products constantly.

A Hilbert space is a vector space with an inner product defined on it. It carries a completeness condition that matters to analysts and rarely to you at this level. Treat it as "the space where states live".

A unitary operator preserves lengths and angles. That is the property that will matter later: it means an operation never changes the total, so whatever normalisation you started with survives. A projection is the opposite kind of thing — it drops a vector onto a subspace and loses whatever pointed elsewhere.

Eigenvectors and eigenvalues answer one question: which vectors does this operator merely stretch, without turning them? Applying the operator to such a vector gives the same vector multiplied by a number. That number is the eigenvalue. Measurement, later, will turn out to be about exactly these.

The number-theory items at the end are here for one reason: they are what Shor's algorithm needs. Greatest common divisor and congruence are the classical machinery surrounding the quantum part of that algorithm. Meeting them now means the algorithm is not two unfamiliar subjects at once.

What you should now be able to explain or do

Build a vector from a basis and say what dimension means. Compute an inner product and say what it measures. Say what an outer product produces, and why confusing it with the inner product breaks later work. Say what a unitary operator preserves and why that matters. Find eigenvalues and eigenvectors, and state what an eigenvector is invariant under. Say why greatest common divisor appears in a quantum course.

Check yourself