A review of white noise analysis from a probabilistic standpoint (Q1365517)

This paper is a review of white noise analysis for probabilist who is not a specialist in this field. The authors start from the idea of Itô integral and its \(S\)-transform. Let \({X_h , h\in C_c^\infty(T)}\), be a real centered Gaussian system with covariance \(\mathbf E[X_h X_f]=\int h(t) f(t) dt.\) The parameter space is extended to \(L^p\) if it is necessary. \(S\)-transform is defined as \(SY(h)=\mathbf E (Y \colon e^{X_h}\colon), Y\in L^p(\Omega)\), where \(\colon e^{X_h}\colon = e^{X_h\frac 1 2 | h| ^2_2}\) is the renormalization. Note that \(SY\) is a non-random (not necessarily linear) functional of \(h\). The \(S\)-transformation of Itô's stochastic integral is \((S\int^t_0 Y_s dB_s)(h)=\int_0^t SY_s(h)h(s)ds\). The authors obtain a new proof of this relation. They also give characterizations of measurability and martingale property in terms of \(S\)-transforms.