Hilbert Spaces: What Are They?

-

Note: this was originally a thread on my twitter. 

 If you want to impress people, you can just say a Hilbert space is just a complete infinite dimensional inner product space and leave it at that, but let's talk about what that actually means. 

When you first learn about vectors, you talk about them as arrows in space; things with a magnitude and a direction. These are elements of \(\mathbb{R}^n\) where n is the number of dimensions of the space you care about. 

A geogebra graph of a vector in three dimensional Euclidean space. 

You also talk about the dot product (or inner product) as a way to tell when vectors are orthogonal. (I'm purposely saying "orthogonal" instead of "perpendicular" here, but when you actually think about arrows, it's the same thing.)

A graph showing the vectors (1,2) and (-2,1) a computation of the dot product of (1,2) and (-2,1) = 0 

 

 As my linear algebra students are about to see, \(\mathbb{R}^n\) is far from the only interesting vector space. A classic example is the space of polynomials of dimension less than or equal to \(n\).

A general polynomial in z 

We know that Taylor series can be used to represent functions that aren't polynomials as "polynomials," albeit infinitely long ones, and this is how I like to motivate infinite dimensional vector spaces.

Our vectors will be infinite sequences, and we'll need some sort of convergence requirement. To make things easy for ourselves, let's use an inner product inspired by \(\mathbb{R}^n\): this is called the \(\ell^2\) inner product. the little l2 inner product 

One thing that's nice in \(\mathbb{R}^n\) is that we have an easy orthonormal basis. We can make "standard basis vectors" for \(\ell^2\) as well. 

a graph showing the standard basis vectors in three dimensional euclidean spacethe k-th standard basis vector for little l2



Remember how in \(\mathbb{R}^n\), you can get the length of the vector by doing the dot product with itself and then taking the square root? That's the metric induced by the inner product. In our infinite dimensional \(\ell^2\), we're going to only consider vectors that have finite length.

taking the euclidean norm of a vector in two dimensional euclidean space 

the little l2 norm 

So far, we've just made an infinite dimensional inner product space. What makes this a Hilbert space is the requirement/property that it is a "complete metric space with the metric induced by the inner product."

A metric space is complete if every Cauchy sequence converges. Think of this as saying "things that get closer and closer together also get closer and closer to a limit point." \(\mathbb{R}^n\) is a complete metric space:

a line with dots that get closer together  

 Sometimes people are disappointed to find out that all finite dimensional vector spaces "are" \(\mathbb{R}^n\), so hopefully I won't disappoint you too much when I say that all (infinite dimensional) Hilbert spaces "are" \(\ell^2\).

More precisely, all *separable* Hilbert spaces are isomorphic to \(\ell^2\). A Hilbert space is separable if it has an orthonormal basis. (All the Hilbert spaces I care about are separable.) This and the completeness requirement save us from uniqueness issues.


"Ok, so if they're all the same, what do you mean when you say 'all the Hilbert spaces?' Doesn't that mean there's only one????" Good question, hypothetical person!

 "Isomorphic" here means that there's a bijective map that preserves inner products. However, this only says the spaces are the same *as Hilbert spaces.* It still allows for different sorts of objects that may have other properties.

Remember we motivated this by talking about Taylor series?

Take a (complex) holomorphic function on the complex unit disk. Being holomorphic means you can write it as a Taylor series that converges in the disk. If you add in a requirement that the coefficients are square summable, it's what we call a Hardy space function.

The Hardy space norm and inner product defined in terms of the coefficients  

Even more exciting, it turns out that there's a way to define the inner product of the Hardy space as a limit of integral means, which gives you the SAME THING.

The Hardy space inner product defined as an integral mean  

The Hilbert space structure tells you some things about this space, but because these are *functions*, there are lots of other properties that come with that. (Factorization, multiplier algebras, boundary behavior, etc)

There are also other ways to define inner products for spaces of functions. These function spaces will be isomorphic (as Hilbert spaces) to \(\ell^2\), but the function theoretic properties will vary.

One way to "tell these apart" is to look at whether or not it's a "Reproducing Kernel Hilbert Space" and to look at what the reproducing kernel is.   

Further reading: 

An Introduction to Hilbert Space by Nicholas Young 

Ryan Tully-Doyle's Hilbert space interactive text: https://rtullydo.github.io/hilbert/frontmatter-1.html

Comments

Post a Comment

Popular posts from this blog

Frostman's Theorem surprises me

-
In my last post , I went into some detail about inner functions, Blaschke products, and singular inner functions. The interesting thing about inner functions that I'll care about here is how they act when you compose them with an automorphism of the disk. This is called a Frostman shift: Frostman shifts continue to be inner functions: inner functions have boundary value modulus 1 and Mobius transforms take the circle to the circle, so the composition also has boundary values modulus 1. What's WILD to me is that for "most" mobius transforms, the Frostman shift of ANY inner function will be a Blaschke product! This surprises me because Blaschke products have zeros in the disk and singular inner functions are zero free in the disk. The proof that Garcia, Mashreghi, and Ross provide is measure theoretic and broadly speaking, it works by showing that the measure for the singular inner part of a Frostman shift is zero (so when integrated and exponentiated, the singular in...
Read more

Two variable shift operators

-
Image
This is my first proper blog post that wasn't just adapted from a twitter thread! Yay! In my last post, I talked about model spaces: the subspaces of the Hardy space \(H^2\) that are shift invariant under the backward shift. One of the things I've been getting into lately is looking at model spaces in two variables, where the situation is more complicated. As a quick refresher, we're living on the Hardy space of the disk, consisting of analytic functions with square summable coefficients. Helpfully, this also corresponds to (radial limits) of integrals around a circle, so we can think of the Hardy space as a subspace of \(L^2(\mathbb{T})\).   Because these functions are analytic, we're thinking about them as a power series, so we can think about multiplication by \(z\) as shifting the coefficients, hence the name "the shift." If we can move coefficients one way, we might want to move them the other way, and this is the backwards shift:   Now, I want to talk ...
Read more

Boundary Values and Inner Functions

-
Image
Someday, I'll do a post that doesn't draw heavily from "Introduction to Model Spaces and their Operators" by Garcia, Mashreghi, and Ross, but today is not that day. I wish I had studied this book as a grad student: it's so well written! Anyway, let's talk about inner functions! First, we need to establish the world in which we live. Today, this is going to be the Banach space of analytic functions bounded on the unit disk, \(H^\infty(\mathbb{D})\). In a complex analysis class, you probably did the exercise where you express the Poisson kernel in a couple of different ways: as a "difference quotient," as the real part of a rational function analytic on the disk, as a geometric series, and in terms of the modulus and argument of the complex variable \(z=re^{i\theta}\): Since the Poisson kernel is the real part of an analytic function, it's harmonic, and taking derivatives, this lets us see that the Poisson integral of a measure \(\mu\) is (comple...
Read more