This note is based on Malliavin Calculus and Related Topics by David Nualart and is intended solely for personal academic study and should not be regarded as a substitute for the original text.
1 Wiener Chaos Decomposition
We seek a generalized form of Gaussian process, say changing the set of time indices
to a "more abstract" space
.
Suppose
is a real separable Hilbert space with innrt product
and norm
for
.
Ito Isometry of High Oder Chaos
In the end of this small section, we consider similar isometry as the Ito isometry that works in
. In the definition of isometric Gaussian process, we have the isometry of oder
as
, i.e.
by the linearity of
and assume
is infinite-dimensional. As we constructed
above, we seek some object that is mapped to
or some multiple of
, which also satisfy the requirements to be an isometry like oder one case:
, or
. Associated with each
, there is a list
, in which
shows up
times (and those
with
don't show up in this list), and the length of this list is
. Moreover, as the number of copies of each base vector
is known, we don't care about their order. The symmetric tensor product in mathematics is such an space whose vectors don't care about order. Consider
, define the symmetric map be
in which
is a permutation. In this end, we can construct the linear map
. Take
such that
, and give the outcome of
acting on all base vectors in symmetric tensor product space
:
Then we consider an appropriate norm on
. A typical choice is the norm
:
Thus we have
To construct an isometry, it is enough to let the norm of
be
while modifying the linear map
to be
This gives an isometry between
and
. Moreover, we can put
all toghether and form so-called Fock space
and thus it is isometric to
.