In this paper we reconsider the existence of stealth fields during the evolution of our Universe by admitting more realistic nonminimal couplings to gravity than the conformal one. In the framework of the FRW cosmology we found that inhomogeneous stealths, as those enabled in the conformal case, are only allowed in de Sitter backgrounds. In particular, it is shown that the homogeneous stealths resulting from nonconformal couplings can coexist with each kind of matter dominant in the several phases characterizing the evolution of our Universe according to the $\Lambda$CDM model.

It is shown that any homogeneous and isotropic universe, independently of its spatial topology and matter content, allows for the presence of a conformal stealth, i.e. a nontrivial conformally invariant scalar field with vanishing energy-momentum tensor, which evolves along with the universe without causing even the smallest backreaction. Surprisingly, this gravitationally invisible universal witness is inhomogeneous with zero consequences for the underlying cosmology. Additionally, it is shown that these results are not exclusive of a four-dimensional universe by generalizing them to higher dimensions.

Robustness of the solutions to the inflaton potential inverse problem based on the slow-roll approximation is addressed. With that aim it is introduced a measure of the difference of the outputs obtained using first and second order respectively in the horizon-flow expansion. The evolution of this measure is determined by a second order linear non-autonomous non-homogeneous differential equation. Boundedness of the general solutions to this equation is analyzed. It is shown that they diverge for most of the physically meaningful cases. Examples for typical inflationary models are presented. It is argued that this lack of robustness is due to the limitations of the slow-roll expansion for probing the scale-dependence of the inflationary spectra.

The observational constraints on the primordial power spectrum have tightened considerably with the release of the first year analysis of the WMAP observations, especially when combined with the results from other CMB experiments and galaxy redshift surveys. These observations allow us to constrain the physics of cosmological inflation: (i) The data show that the Hubble distance is almost constant during inflation. While observable modes cross the Hubble scale, it changes by less than 3% during one e-folding: d(d_H)/dt < 0.032 at 2 sigma. The distance scale of inflation itself remains poorly constrained: 1.2 x 10^{-28} cm < d_H < 1 cm. (ii) We present a new classification of single-field inflationary scenarios (including scenarios beyond slow-roll inflation), based on physical criteria, namely the behaviour of the kinetic and total energy densities of the inflaton field. The current data show no preference for any of the scenarios. (iii) For the first time the slow-roll assumption could be dropped from the data analysis and replaced by the more general assumption that the Hubble scale is (almost) constant during the observable part of inflation. We present simple analytic expressions for the scalar and tensor power spectra for this very general class of inflation models and test their accuracy.

Conditions for accelerated expansion of Friedmann-Robertson-Walker space-time are analyzed. Connection of this scenario with present-day observations are reviewed. It is explained how a scalar field could be responsible for cosmic acceleration observed in present times and predicted for the very early Universe. Ideas aimed at answering whether is that the actual case for our Universe are described.

For the inflaton perturbations it is shown that the evolution of the difference between the spectral indices can be translated into information on the scale dependence of the tensor to scalar amplitudes ratio, $r$, and how the scalar field potential can be derived from that information. Examples are given where $r$ converges to a constant value during inflation but dynamics are rather different from the power--law model. Cases are presented where a constant $r$ is not characteristic of the inflationary dynamics though the resulting perturbation spectra are consistent with the CMB and LSS data. The inflaton potential corresponding to $r$ given by a n--th order polynomial of the e--folds number is derived in quadratures expressions. Since the observable difference between the spectral indices evaluated at a pivot scale yields information about the linear term of that polynomial, the first order case is explicitly written down. The solutions show features beyond the exponential form corresponding to power--law inflation and can be matched with current observational data.

It is argued that the order of the analytic expressions for the calculation of the primordial perturbations from inflation exerts a strong influence upon the results of the analysis of observables dynamics based on these expressions and, therefore, upon some predictions sometimes taken for granted as generic for the inflationary scenario.

Based on solutions of the Stewart-Lyth inverse problem it is argued that in the CMB data analysis the parametrization of the primordial spectra from inflation must include the `running' of both, scalar and tensorial, spectral indices if information beyond the exponential potential model is wanted to be detected.

The single scalar field inflationary models that lead to scalar and tensor perturbation spectra with amplitudes varying in direct proportion to one another are reconstructed by solving the Stewart-Lyth inverse problem to next-to-leading order in the slow-roll approximation. The potentials asymptote at high energies to an exponential form, corresponding to power law inflation, but diverge from this model at low energies, indicating that power law inflation is a repellor in this case. This feature implies that a fine-tuning of initial conditions is required if such models are to reproduce the observations. The required initial conditions might be set through the eternal inflation mechanism. If this is the case, it will imply that the spectral indices must be nearly constant, making the underlying model observationally indistinguishable from power law inflation.

Next-to-leading order expressions related to Stewart-Lyth inverse problem are used to determine the inflationary models with a tensorial power spectrum described by a scale-invariant spectral index. Beyond power-law inflation, solutions are characterized by scale dependent scalar indices. These models can be used as assumption on the generation of primordial perturbations to test for scale dependence of the scalar index at large angular scales. If such a dependence is detected, a nonzero contribution of gravitational waves to the CMB spectrum must be expected.

Recently it was argued that an inflationary potential yielding power spectra characterized by a scale-invariant tensorial spectral index and a weakly scale-dependent scalar spectral index might account for a generic large-scale structure formation in the multiverses scenario of eternal inflation. Here it is shown that this statement remains true if the tensorial index is allowed to slowly vary in a wide range of angular scales. In the cases analyzed in this paper, the scalar index also slightly depends on the scales. Therefore, the production of closely resembling each other universes seems to be a common feature of inflationary scenarios with weakly scale-dependent spectral indices. With power-law inflation as the extreme case, this is the class of scenarios which may give the best fit to near future CMB observations.

We calculate power spectra of cosmological perturbations at high accuracy for two classes of inflation models. We classify the models according to the behaviour of the Hubble distance during inflation. Our approximation works if the Hubble distance can be approximated either to be a constant or to grow linearly with cosmic time. Many popular inflationary models can be described in this way, e.g., chaotic inflation with a monomial potential, power-law inflation and inflation at a maximum. Our scheme of approximation does not rely on a slow-roll expansion. Thus we can make accurate predictions for some of the models with large slow-roll parameters.

Recently an inflationary potential yielding power spectra characterized by a scale-invariant tensorial spectral index and a scale-dependent scalar spectral index was introduced. We analize here the implications that this potential could have for the large-scale structure formation in the multiverses scenario of eternal inflation.

We present a new approximation scheme that allows us to increase the accuracy of analytical predictions of the power spectra of inflationary perturbations for two specific classes of inflationary models. Among these models are chaotic inflation with a monomial potential, power-law inflation and natural inflation (inflation at a maximum). After reviewing the established first order results we calculate the amplitudes and spectral indices for these classes of models at higher orders in the slow-roll parameters for scalar and tensorial perturbations.

In this paper the Stewart-Lyth inverse problem is introduced. It consists of solving two non-linear differential equations for the first slow-roll parameter and finding the inflaton potential. The equations are derived from the Stewart--Lyth equations for the scalar and tensorial perturbations produced during the inflationary period. The geometry of the phase planes transverse to the trajectories is analyzed, and conclusions about the possible behaviour for general solutions are drawn.

In this paper the Stewart-Lyth inverse problem is rewritten using the comoving scales as the basic parameter. It is shown that some information on the inflaton potential can be obtained from observations taking into account only the scalar power spectrum.