1: \begin{abstract}

2: This is an overview about   natural  sample spaces for

3: differential equations driven by various noises. Appropriate

4: sample spaces are needed in order to facilitate a random dynamical

5: systems approach for stochastic differential equations. The noise

6: could be white or colored, Gaussian or non-Gaussian, Markov or

7: non-Markov, and semimartingale or non-semimartingale.  Typical

8: noises are defined in terms of Brownian motion, L\'evy motion and

9: fractional Brownian motion. In each of these cases, a canonical

10: sample space with an appropriate metric (or topology that gives

11: convergence concept) is introduced. Basic properties of

12: canonical sample spaces, such as separability and completeness,

13: are then discussed.

14: 

15: Moreover, a flow defined by shifts, is introduced on these

16: canonical sample spaces. This flow has an invariant measure which

17: is the probability distribution for Brownian motion, or L\'evy

18: motion or fractional Brownian motion.

19: % Two equivalent definitions

20: %of Brownian motion, the Wiener (probability) measure, L\'evy

21: %(probability) measure, and the L\'evy  (jump) measure are

22: %reviewed. This is done in the context of canonical sample spaces

23: %for stochastic differential equations driven by Gaussian Brownian

24: %motion or non-Gaussian L\'evy motion.

25: Thus canonical sample spaces are much richer in mathematical

26: structures than the usual sample spaces in probability theory, as

27: they have metric or topological structures, together with a

28:   shift flow (or driving flow) defined on it.  This

29: facilitates dynamical systems approaches for studying stochastic

30: differential equations.

31: 

32: 

33: 

34: \end{abstract}

35: