No Arabic abstract
The necessary information to distinguish a local inhomogeneous mass density field from its spatial average on a compact domain of the universe can be measured by relative information entropy. The Kullback-Leibler (KL) formula arises very naturally in this context, however, it provides a very complicated way to compute the mutual information between spatially separated but causally connected regions of the universe in a realistic, inhomogeneous model. To circumvent this issue, by considering a parametric extension of the KL measure, we develop a simple model to describe the mutual information which is entangled via the gravitational field equations. We show that the Tsallis relative entropy can be a good approximation in the case of small inhomogeneities, and for measuring the independent relative information inside the domain, we propose the Renyi relative entropy formula.
I show that a generic quantum phenomenon can drive cosmic acceleration without the need for dark energy or modified gravity. When treating the universe as a quantum system, one typically focuses on the scale factor (of an FRW spacetime) and ignores many other degrees of freedom. However, the information capacity of the discarded variables will inevitably change as the universe expands, generating quantum bias (QB) in the Friedmann equations [Phys. Lett. A 382, 36, 2555 (2018)|arXiv:1707.05789]. If information could be stored in each Planck-volume independently, this effect would give rise to a constant acceleration $10^{120}$ times larger than that observed, reproducing the usual cosmological constant problem. However, once information capacity is quantified according to the holographic principle, cosmic acceleration is far smaller and depends on the past behaviour of the scale factor. I calculate this holographic quantum bias, derive the semiclassical Friedmann equations, and obtain their general solution for a spatially-flat universe containing matter and radiation. Comparing these QB-CDM solutions to those of $Lambda$CDM, the new theory is shown to be falsifiable, but nonetheless consistent with current observations. In general, realistic QB cosmologies undergo phantom acceleration ($w_mathrm{eff}<-1$) at late times, predicting a Big Rip in the distant future.
We investigate the analogy between the renormalization group (RG) and deep neural networks, wherein subsequent layers of neurons are analogous to successive steps along the RG. In particular, we quantify the flow of information by explicitly computing the relative entropy or Kullback-Leibler divergence in both the one- and two-dimensional Ising models under decimation RG, as well as in a feedforward neural network as a function of depth. We observe qualitatively identical behavior characterized by the monotonic increase to a parameter-dependent asymptotic value. On the quantum field theory side, the monotonic increase confirms the connection between the relative entropy and the c-theorem. For the neural networks, the asymptotic behavior may have implications for various information maximization methods in machine learning, as well as for disentangling compactness and generalizability. Furthermore, while both the two-dimensional Ising model and the random neural networks we consider exhibit non-trivial critical points, the relative entropy appears insensitive to the phase structure of either system. In this sense, more refined probes are required in order to fully elucidate the flow of information in these models.
Classical chaos is often characterized as exponential divergence of nearby trajectories. In many interesting cases these trajectories can be identified with geodesic curves. We define here the entropy by $S = ln chi (x)$ with $chi(x)$ being the distance between two nearby geodesics. We derive an equation for the entropy which by transformation to a Ricatti-type equation becomes similar to the Jacobi equation. We further show that the geodesic equation for a null geodesic in a double warped space time leads to the same entropy equation. By applying a Robertson-Walker metric for a flat three-dimensional Euclidian space expanding as a function of time, we again reach the entropy equation stressing the connection between the chosen entropy measure and time. We finally turn to the Raychaudhuri equation for expansion, which also is a Ricatti equation similar to the transformed entropy equation. Those Ricatti-type equations have solutions of the same form as the Jacobi equation. The Raychaudhuri equation can be transformed to a harmonic oscillator equation, and it has been shown that the geodesic deviation equation of Jacobi is essentially equivalent to that of a harmonic oscillator. The Raychaudhuri equations are strong geometrical tools in the study of General Relativity and Cosmology. We suggest a refined entropy measure applicable in Cosmology and defined by the average deviation of the geodesics in a congruence.
We study the emergence of entropy in gravitational production of dark matter particles, ultra light scalars minimally coupled to gravity and heavier fermions, from inflation to radiation domination (RD). Initial conditions correspond to dark matter fields in their Bunch-Davies vacua during inflation. The out states are correlated particle-antiparticle pairs, and the distribution function is found in both cases. In the adiabatic regime the density matrix features rapid decoherence by dephasing from interference effects in the basis of out particle states, effectively reducing it to a diagonal form with a concomitant von Neumann entropy. We show that it is exactly the entanglement entropy obtained by tracing over one member of the correlated pairs. Remarkably, for both statistics the entanglement entropy is similar to the quantum kinetic entropy in terms of the distribution function with noteworthy differences stemming from pair correlations. The entropy and the kinetic fluid form of the energy momentum tensor all originate from decoherence of the density matrix. For ultra light scalar dark matter, the distribution function peaks at low momentum $propto 1/k^3$ and the specific entropy is $ll 1$. This is a hallmark of a emph{condensed phase} but with vanishing field expectation value. For fermionic dark matter the distribution function is nearly thermal and the specific entropy is $mathcal{O}(1)$ typical of a thermal species. We argue that the functional form of the entanglement entropy is quite general and applies to alternative production mechanisms such as parametric amplification during reheating.
We introduce an axiomatic approach to entropies and relative entropies that relies only on minimal information-theoretic axioms, namely monotonicity under mixing and data-processing as well as additivity for product distributions. We find that these axioms induce sufficient structure to establish continuity in the interior of the probability simplex and meaningful upper and lower bounds, e.g., we find that every relative entropy must lie between the Renyi divergences of order $0$ and $infty$. We further show simple conditions for positive definiteness of such relative entropies and a characterisation in term of a variant of relative trumping. Our main result is a one-to-one correspondence between entropies and relative entropies.