1: \begin{abstract} {We develop connections between Stein's approximation

2: method, logarithmic Sobolev and transport inequalities by

3: introducing a new class of functional inequalities involving the relative entropy,

4: the Stein kernel, the relative Fisher information and the Wasserstein distance with respect to

5: a given reference distribution on $\R^d$. For the Gaussian model,

6: the results improve upon the classical logarithmic Sobolev inequality and

7: the Talagrand quadratic transportation cost

8: inequality. Further examples of illustrations include

9: multidimensional gamma distributions, beta distributions,

10: as well as families of log-concave densities.

11: As a by-product, the new inequalities are shown to be relevant

12: towards convergence to equilibrium, concentration inequalities and entropic convergence expressed

13: in terms of the Stein kernel. The tools rely on semigroup interpolation

14: and bounds, in particular by means of the iterated gradients of the Markov generator

15: with invariant measure the distribution under consideration.

16: In a second part, motivated by the

17: recent investigation by Nourdin, Peccati and Swan on Wiener chaoses, we address the issue

18: of entropic bounds on multidimensional functionals $F$ with the Stein kernel via

19: a set of data on $F$ and its gradients rather than on the Fisher information of the density.

20: A natural framework for this investigation is given by the Markov Triple structure

21: $(E, \mu, \Gamma)$ in which abstract Malliavin-type arguments may be developed

22: and extend the Wiener chaos setting. } \\

23: 

24: 

25: 

26: 

27: 

28: \noindent \textbf{Keywords:} Entropy, Fisher Information, Stein Kernel

29: and Discrepancy, Logarithmic Sobolev Inequality, Transport Inequality,

30: Convergence to Equilibrium, Concentration Inequality,

31: Normal Approximation, Gamma Calculus.\\

32: \\

33:  \noindent

34:  \textbf{2000 Mathematics Subject Classification:} 60E15, 26D10, 60B10

35: \end{abstract}