7
$\begingroup$

Recall that a Hilbert space $\mathcal{H}$ is a reproducing kernel Hilbert space (RKHS) if the elements of $\mathcal{H}$ are functions on a certain set $X$ and for any $a\in X$, the linear functional $f\mapsto f(a)$ is bounded on $\mathcal{H}$. By Riesz Representation Theorem, there exists an element $K_a\in\mathcal{H}$ such that $$f(a) = \langle f, K_a\rangle\ \text{ for all } \ f\in\mathcal{H}.$$ The function $K(x,y) = K_y(x) = \langle K_y, K_x\rangle$ defined on $X\times X$ is called the reproducing kernel function of $\mathcal{H}$.

It is well known and easy to show that for any orthonormal basis $\{e_m\}_{m=1}^{\infty}$ for $\mathcal{H}$, we have the formula $$K(x,y) = \sum_{m=1}^{\infty}e_m(x)\overline{e_m(y)},\tag{Eqn 1}$$ where the convergence is pointwise on $X\times X$.

My question concerns the converse of the above statement.

Question: if $\{g_m\}_{m=1}^{\infty}$ is a sequence of functions in $\mathcal{H}$ such that $$K(x,y) = \sum_{m=1}^{\infty}g_m(x)\overline{g_m(y)}\tag{Eqn 2}$$ for all $x,y\in X$. Is the sequence $\{g_m\}_{m=1}^{\infty}$ an orthonormal basis for $\mathcal{H}$?

The answer to this question is clearly negative since equation (Eqn 1) can be re-written as $$K(x,y) = \frac{e_1(x)}{\sqrt{2}}\overline{\frac{e_1(y)}{\sqrt{2}}}+\frac{e_1(x)}{\sqrt{2}}\overline{\frac{e_1(y)}{\sqrt{2}}}+\sum_{m=2}^{\infty}e_m(x)\overline{e_m(y)}$$ and clearly $\{e_1/\sqrt{2}, e_1/\sqrt{2}, e_2, \ldots\}$ is not an orthonormal basis for $\mathcal{H}$. So the following additional condition should be added: the sequence $\{g_m\}_{m=1}^{\infty}$ is linearly independent.

The following proof suggests that the answer is affirmative. (For those who are familiar with the proof of the Moore-Aronszajn's Theorem in the theory of RKHS, the proof here looks similar.) Assume that we have (Eqn 2) and the sequence $\{g_m\}_{m=1}^{\infty}$ is linearly independent. Let $\mathcal M$ be the linear space spanned by the functions $\{g_m\}_{m=1}^{\infty}$. Define a sesquilinear form on $\mathcal M$ as \begin{align*} \left\langle\sum_{\text{finite sum}}a_jg_j, \sum_{\text{finite sum}}b_kg_k\right\rangle_{\mathcal M} = \sum_{\text{finite sum}} a_j\overline{b}_j. \end{align*} Since $\{g_m\}_{m=1}^{\infty}$ is a linearly independent set, the above definition is well-defined. Note that $\{g_m\}_{m=1}^{\infty}$ is an orthonormal set in $\langle,\rangle_{\mathcal M}$. For any $f\in\mathcal M$ and $x\in X$, we have \begin{align*} f(x) = \sum_{\text{finite sum}}\langle f,g_m\rangle_{\mathcal M}\,g_m(x). \end{align*} Cauchy-Schwarz's inequality gives \begin{align*} |f(x)| & \leq \Big(\sum_{\text{finite sum}}|\langle f,g_m\rangle_{\mathcal M}|^2\Big)^{1/2}\Big(\sum_{\text{finite sum}}|g_m(x)|^2\Big)^{1/2} \leq \|f\|_{\mathcal M}\sqrt{K(x,x)}. \end{align*} Let $\widetilde{\mathcal M}$ be the Hilbert space completion of $\mathcal M$. The standard argument shows that $\widetilde{\mathcal M}$ is a RKHS of functions on $X$. What is the kernel of $\widetilde{\mathcal M}$? Since $\{g_m\}_{m=1}^{\infty}$ is an orthonormal set and its span is dense in $\widetilde{\mathcal M}$, it is an orthonormal basis for $\widetilde{\mathcal M}$. The kernel of $\widetilde{\mathcal M}$ then can be computed as $$\sum_{m=1}^{\infty}g_m(x)\bar{g}_m(y),$$ which is the same as $K(x,y)$. Therefore, $\widetilde{\mathcal M}$ is the same as $\mathcal H$ (they consist of the same functions and the inner products on the two spaces are equal). Consequently, $\{g_m\}_{m=1}^{\infty}$ is an orthonormal basis for $\mathcal{H}$. This completes the proof.

Counterexample: On the other hand, there are counterexamples that provide a negative answer to the question in the infinite dimensional case.

What part of the above proof is incorrect? I have checked but could not figure out what went wrong.

Thank you.

$\endgroup$
9
  • 1
    $\begingroup$ Could you describe briefly one of these counterexamples? $\endgroup$ Jan 23, 2016 at 19:43
  • 1
    $\begingroup$ @ChrisRamsey: Together with Theorem 3.12 in Vern Paulsen's lecture note at www.math.uh.edu/~vern/rkhs.pdf, if the conclusion were correct, then any linearly independent Parseval frame would be an orthonormal basis. However this seems to be incorrect. For example, take E to be the orthogonal complement of v = (1, 1/2, 1/3, ...) in l^2 and take g_m = P(e_m), where P is the orthogonal projection onto E. Then {g_m} is a linearly independent Parseval frame for E but it is not an orthonormal basis. $\endgroup$
    – T. Le
    Jan 23, 2016 at 19:59
  • 1
    $\begingroup$ This sentence worries me: "The standard argument shows that $\widetilde{\mathcal{M}}$ is a RKHS of functions on $X$". Can you look at that more carefully? I am not convinced that the completion $\widetilde{\mathcal{M}}$ can be identified with a space of functions on $X$. You have shown that the map $f \mapsto f(x)$ is continuous with respect to the $\mathcal{M}$ norm, so this map extends to the completion, and every element $\tilde{f}$ of the completion is associated with a well-defined function on $X$. But I do not see why this association has to be 1-1. $\endgroup$ Jan 23, 2016 at 20:38
  • $\begingroup$ By "standard argument", I meant the usual argument that one might find when proving that any positive definite kernel K(x,y) gives rise to an RKHS whose kernel is K(x,y) (Moore-Aronszajn Theorem). Such argument can be found in Vern Paulsen's note, page 14, at www.math.uh.edu/~vern/rkhs.pdf or in the proof of Theorem 2.23 on page 19 in Agler-McCarthy's book "Pick Interpolation and Hilbert Function Spaces". @NateEldredge: can you explain what you mean and why you need the association to be 1-1 for $\widetilde{\mathcal{M}}$ to be RKHS? $\endgroup$
    – T. Le
    Jan 23, 2016 at 23:02
  • 1
    $\begingroup$ @NateEldredge: I think you are right! I was unable to verify that the map from $\widetilde{\mathcal{M}}$ to $\mathbb{R}^{X}$ (or $\mathbb{C}^{X}$ for the complex case) is injective. In other words, if $h\in\widetilde{\mathcal{M}}$ and $h(x)=0$ for all $x\in X$, then it cannot be verified that $\|h\|=0$ (in fact, this may not be true). On the other hand, in the "standard proofs" of Moore-Aronszajn Theorem, injectivity can be verified (even though in all the proofs that I know of, it was not mentioned or verified!!). Could you please put part of your comments in an answer so I can vote for it? $\endgroup$
    – T. Le
    Jan 24, 2016 at 1:13

1 Answer 1

8
$\begingroup$

The error is in this line:

The standard argument shows that $\widetilde{\mathcal{M}}$ is an RKHS of functions on $X$.

In fact, this is not generally true. The completion $\widetilde{\mathcal{M}}$ may not be naturally identified with a space of functions on $X$.

The "obvious" way that one would try to prove this is as follows. Consider an element $\phi \in \widetilde{\mathcal{M}}$. Since $\mathcal{M}$ is dense in its completion, there is a sequence $\{f_n\} \subset \mathcal{M}$ such that $f_n \to \phi$ in $\widetilde{\mathcal{M}}$-norm, which is an extension of the $\mathcal{M}$-norm. In particular, $\{f_n\}$ is $\mathcal{M}$-norm Cauchy. Because of your inequality $|f(x)| \le \|f\|_{\mathcal{M}} \sqrt{K(x,x)}$ (*), we have that $\{f_n(x)\}$ is Cauchy in $\mathbb{R}$ for each $x$. So $\{f_n(x)\}$ converges to a number which we may call $f_\phi(x)$; that is, $f_n \to f_\phi$ pointwise. It is also easy to show that $f_{\phi}$ does not depend on the choice of sequence $f_n \to \phi$, so the linear map $\phi \mapsto f_\phi$ of $\widetilde{\mathcal{M}}$ into $\mathbb{R}^X$ is well defined. Let's call this map $T$.

So the "obvious" thing to do is to look at the image of $T$, which is of course a function space, and make it a Hilbert space by pushing forward the $\widetilde{\mathcal{M}}$-norm. If this works, then the same inequality (*) will show that the evaluation map is continuous in this norm, and we have ourselves an RKHS.

The problem is that $T$ might fail to be injective. In other words, we could have a nonzero $\phi$ for which $f_\phi = 0$. In that case, pushing forward the norm will not work; it will not be well defined on the image of $T$.

This is exactly what happens in the counterexample you discuss in the comments. As you suggest, consider $\ell^2$ as an RKHS on $X = \mathbb{N} = \{1,2,\dots\}$, with its usual orthonormal basis $\{e_n\}$, and let $v = \sum_n \frac{1}{n} e_n$. Let $H = \{v\}^\perp$ with the same $\ell^2$ inner product; being a closed subspace of $\ell^2$, $H$ is an RKHS.

Let $P : \ell^2 \to H$ be the orthogonal projection and let $g_n = P e_n$. The $g_n$ are linearly independent; if $0 = a_1 g_1 + \dots + a_n g_n = P(a_1 e_1 + \dots + a_n e_n)$ then $a_1 e_1 + \dots + a_n e_n$ is a scalar multiple of $v$, which is only possible if all $a_i$ are 0.

Following your construction, let $\mathcal{M} \subset H$ be the linear span of $\{g_n\}$ and let $\langle\cdot, \cdot\rangle_{\mathcal{M}}$ be the inner product on $\mathcal{M}$ which makes the $\{g_n\}$ orthonormal.

Set $h_m = \sum_{n=1}^m \frac{1}{n} g_n$. Clearly the sequence $\{h_m\}$ is Cauchy in $\mathcal{M}$, so it converges in the completion $\widetilde{\mathcal{M}}$ to some $\phi$. It is also clear that $\|\phi\|_\widetilde{\mathcal{M}}^2 = \frac{\pi^2}{6}$ so in particular $\phi \ne 0$.

But on the other hand, we have $h_m = \sum_{n=1}^m \frac{1}{n} P e_n = P\left(\sum_{n=1}^m \frac{1}{n} e_n\right)$. By continuity of $P$, we have $h_m \to Pv = 0$ in $\ell^2$, and thus also pointwise. So $f_\phi = 0$.


This is basically an example of a somewhat paradoxical fact that's bitten me before. Let $X,Y$ be Banach spaces, and suppose $E \subset X$ is dense. It's well known that every bounded operator $T : E \to Y$ has a unique bounded extension $\tilde{T} : X \to Y$. But it's possible that $T$ is injective while $\tilde{T}$ is not. Indeed, we've just constructed an example, by letting $E = \mathcal{M}$, $X = \widetilde{\mathcal{M}}$, $Y = \ell^2$, and $T : E \to Y$ the inclusion map.

$\endgroup$
3
  • $\begingroup$ Nice answer. Just a remark related to your last point: we get particularly striking examples if X is the full group Cstar algebra of a non-amenable discrete group (like $F_2$), E is the dense subalgebra generated by point masses, and Y is the reduced group Cstar algebra. Then the "identity map" from E to Y is continuous and injective with dense range, but the unique continuous extension to a map $X \to Y$ is not injective $\endgroup$
    – Yemon Choi
    Jan 24, 2016 at 3:25
  • $\begingroup$ Thank you for your answer! On the other hand, I believe that the map $T$ is injective and hence the Question has an affirmative answer if we assume a stronger condition than just linear independence: for any sequence $\{c_m\}$ in $\ell^2$, if $\sum_{m=1}^{\infty}c_mg_m(x) = 0$ for all $x\in X$, then $c_1=c_2=\cdots = 0$ (note that the convergence of the series is automatic by Cauchy-Schwarz's inequality). I think there is a name for this condition but I'm not sure what it is. $\endgroup$
    – T. Le
    Jan 24, 2016 at 3:46
  • 1
    $\begingroup$ @T.Le: Yes, that's true, because the map $S : \ell^2 \to \widetilde{\mathcal{M}}$ which maps $\{c_m\}$ to $ \sum_m c_m g_m$ is an isometric isomorphism. Your condition is that $TS$ is injective, which means that $T$ must be as well. $\endgroup$ Jan 24, 2016 at 3:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.