-
WKB Approximation for the Polymer Quantization of the Taub Model
Authors:
Valerio Cascioli,
Giovanni Montani,
Riccardo Moriconi
Abstract:
We develop a suitable technical algorithm to implement a separation of the Minisuperspace configurational variables into quasi-classical and purely quantum degrees of freedom, in the framework of a Polymer quantum Mechanics reformulation of the canonical dynamics. We then implement the obtained general scheme to the specific case of a Taub Universe, in the presence of a free massless scalar field.…
▽ More
We develop a suitable technical algorithm to implement a separation of the Minisuperspace configurational variables into quasi-classical and purely quantum degrees of freedom, in the framework of a Polymer quantum Mechanics reformulation of the canonical dynamics. We then implement the obtained general scheme to the specific case of a Taub Universe, in the presence of a free massless scalar field. In particular, we identify the quasi-classical variables in the Universe volume and a suitable function of the scalar field, while the purely quantum degree of freedom corresponds to the Universe anisotropy. We demonstrate that the Taub cosmology is associated to a cyclical Universe, oscillating between a minimum and maximum volume turning points, respectively. The pure quantum Universe anisotropy has always a finite value.
△ Less
Submitted 1 June, 2020; v1 submitted 22 March, 2019;
originally announced March 2019.
-
High-dimensional Bayesian optimization using low-dimensional feature spaces
Authors:
Riccardo Moriconi,
Marc P. Deisenroth,
K. S. Sesh Kumar
Abstract:
Bayesian optimization (BO) is a powerful approach for seeking the global optimum of expensive black-box functions and has proven successful for fine tuning hyper-parameters of machine learning models. However, BO is practically limited to optimizing 10--20 parameters. To scale BO to high dimensions, we usually make structural assumptions on the decomposition of the objective and\slash or exploit t…
▽ More
Bayesian optimization (BO) is a powerful approach for seeking the global optimum of expensive black-box functions and has proven successful for fine tuning hyper-parameters of machine learning models. However, BO is practically limited to optimizing 10--20 parameters. To scale BO to high dimensions, we usually make structural assumptions on the decomposition of the objective and\slash or exploit the intrinsic lower dimensionality of the problem, e.g. by using linear projections. We could achieve a higher compression rate with nonlinear projections, but learning these nonlinear embeddings typically requires much data. This contradicts the BO objective of a relatively small evaluation budget. To address this challenge, we propose to learn a low-dimensional feature space jointly with (a) the response surface and (b) a reconstruction map**. Our approach allows for optimization of BO's acquisition function in the lower-dimensional subspace, which significantly simplifies the optimization problem. We reconstruct the original parameter space from the lower-dimensional subspace for evaluating the black-box function. For meaningful exploration, we solve a constrained optimization problem.
△ Less
Submitted 25 September, 2020; v1 submitted 27 February, 2019;
originally announced February 2019.
-
Bianchi I model as a prototype for a cyclical Universe
Authors:
Giovanni Montani,
Andrea Marchi,
Riccardo Moriconi
Abstract:
We analyze the dynamics of the Bianchi I model in the presence of stiff matter, an ultrarelativistic component and a small negative cosmological constant. We quantize this model in the framework of the polymer quantum mechanics, in order to introduce cut-off features in the minisuperspace dynamics.
We then apply to the polymer Wheeler-DeWitt equation, emerging from the Dirac constraint, an adiab…
▽ More
We analyze the dynamics of the Bianchi I model in the presence of stiff matter, an ultrarelativistic component and a small negative cosmological constant. We quantize this model in the framework of the polymer quantum mechanics, in order to introduce cut-off features in the minisuperspace dynamics.
We then apply to the polymer Wheeler-DeWitt equation, emerging from the Dirac constraint, an adiabatic approximation a la Vilenkin, which treats the Universe volume as a quasi-classical variable, becoming de facto the dynamical clock for the pure quantum degrees of freedom, here identified in the Universe anisotropies.
The main issue of the present analysis consists of determining a cyclical evolution for the Bianchi I model, oscillating between the Big-Bounce induced by the cut-off physics and the turning point due to the small cosmological constant. Furthermore, the mean value of the Universe anisotropy variables remains finite during the whole evolution, including the phase across the Big-Bounce. Such a feature, according to a suitable choice of the initial conditions makes the present cosmological paradigm, a viable scenario for the description of a possible primordial and late phases of the actual Universe.
△ Less
Submitted 15 May, 2018;
originally announced May 2018.
-
The reparameterization trick for acquisition functions
Authors:
James T. Wilson,
Riccardo Moriconi,
Frank Hutter,
Marc Peter Deisenroth
Abstract:
Bayesian optimization is a sample-efficient approach to solving global optimization problems. Along with a surrogate model, this approach relies on theoretically motivated value heuristics (acquisition functions) to guide the search process. Maximizing acquisition functions yields the best performance; unfortunately, this ideal is difficult to achieve since optimizing acquisition functions per se…
▽ More
Bayesian optimization is a sample-efficient approach to solving global optimization problems. Along with a surrogate model, this approach relies on theoretically motivated value heuristics (acquisition functions) to guide the search process. Maximizing acquisition functions yields the best performance; unfortunately, this ideal is difficult to achieve since optimizing acquisition functions per se is frequently non-trivial. This statement is especially true in the parallel setting, where acquisition functions are routinely non-convex, high-dimensional, and intractable. Here, we demonstrate how many popular acquisition functions can be formulated as Gaussian integrals amenable to the reparameterization trick and, ensuingly, gradient-based optimization. Further, we use this reparameterized representation to derive an efficient Monte Carlo estimator for the upper confidence bound acquisition function in the context of parallel selection.
△ Less
Submitted 1 December, 2017;
originally announced December 2017.
-
Behavior of the Universe anisotropy in a big-bounce cosmology
Authors:
Riccardo Moriconi,
Giovanni Montani
Abstract:
We investigate the classical and quantum behavior of a Bianchi I model in the presence of a stiff matter contribution when the Vilenkin interpretation of the wave function of the Universe is taken into account. We study its evolution in the so-called polymer representation of quantum mechanics, in order to characterize the modifications that a discrete nature in the isotropic variable of the Unive…
▽ More
We investigate the classical and quantum behavior of a Bianchi I model in the presence of a stiff matter contribution when the Vilenkin interpretation of the wave function of the Universe is taken into account. We study its evolution in the so-called polymer representation of quantum mechanics, in order to characterize the modifications that a discrete nature in the isotropic variable of the Universe induces on the morphology of the cosmological singularity. We demonstrate that in such a model the Big-Bang singularity is removed at a semiclassical level in favor of a Big-Bounce when a lattice on the isotropic variable is considered. Furthermore, the analysis of the mean values on the quantum degrees of freedom, \textit{i.e} the variables $β_{+},β_{-}$ in the Misner picture, and the investigation on the evolution of the wave packets show how the typical diverging behavior associated to the anisotropies of the Universe in proximity of the initial singularity disappears in our polymer modified scheme. Indeed, the anisotropies remain finite across the Big-Bounce and they assume a value that depends on the initial conditions fixed far from the turning point. Finally, we demonstrate that the proposed scenario can be extended, with a suitable choice of the configuration parameters, to the Bianchi IX cosmology and therefore it can be regarded as a paradigm for the generic cosmological model.
△ Less
Submitted 14 June, 2017;
originally announced June 2017.
-
Big-bounce cosmology from quantum gravity: The case of a cyclical Bianchi I universe
Authors:
Riccardo Moriconi,
Giovanni Montani,
Salvatore Capozziello
Abstract:
We analyse the classical and quantum dynamics of a Bianchi I model in the presence of a small negative cosmological constant characterizing its evolution in term of the dust-time dualism. We demonstrate that in a canonical metric approach, the cosmological singularity is removed in correspondence to a positive defined value of the dust energy density. Furthermore, the quantum Big-Bounce is connect…
▽ More
We analyse the classical and quantum dynamics of a Bianchi I model in the presence of a small negative cosmological constant characterizing its evolution in term of the dust-time dualism. We demonstrate that in a canonical metric approach, the cosmological singularity is removed in correspondence to a positive defined value of the dust energy density. Furthermore, the quantum Big-Bounce is connected to the Universe turning point via a well-defined semiclassical limit. Then we can reliably infer that the proposed scenario is compatible with a cyclical Universe picture. We also show how, when the contribution of the dust energy density is sufficiently high, the proposed scenario can be extended to the Bianchi IX cosmology and therefore how it can be regarded as a paradigm for the generic cosmological model. Finally, we investigate the origin of the observed cut-off on the cosmological dynamics, demonstrating how the Big-Bounce evolution can be mimicked by the same semiclassical scenario, where the negative cosmological constant is replaced via a polymer discretization of the Universe volume. A direct proportionality law between such two parameters is then established.
△ Less
Submitted 1 September, 2016; v1 submitted 15 July, 2016;
originally announced July 2016.
-
Safe Learning of Regions of Attraction for Uncertain, Nonlinear Systems with Gaussian Processes
Authors:
Felix Berkenkamp,
Riccardo Moriconi,
Angela P. Schoellig,
Andreas Krause
Abstract:
Control theory can provide useful insights into the properties of controlled, dynamic systems. One important property of nonlinear systems is the region of attraction (ROA), a safe subset of the state space in which a given controller renders an equilibrium point asymptotically stable. The ROA is typically estimated based on a model of the system. However, since models are only an approximation of…
▽ More
Control theory can provide useful insights into the properties of controlled, dynamic systems. One important property of nonlinear systems is the region of attraction (ROA), a safe subset of the state space in which a given controller renders an equilibrium point asymptotically stable. The ROA is typically estimated based on a model of the system. However, since models are only an approximation of the real world, the resulting estimated safe region can contain states outside the ROA of the real system. This is not acceptable in safety-critical applications. In this paper, we consider an approach that learns the ROA from experiments on a real system, without ever leaving the true ROA and, thus, without risking safety-critical failures. Based on regularity assumptions on the model errors in terms of a Gaussian process prior, we use an underlying Lyapunov function in order to determine a region in which an equilibrium point is asymptotically stable with high probability. Moreover, we provide an algorithm to actively and safely explore the state space in order to expand the ROA estimate. We demonstrate the effectiveness of this method in simulation.
△ Less
Submitted 16 August, 2017; v1 submitted 15 March, 2016;
originally announced March 2016.
-
Chaos Removal in the $R+qR^2$ gravity: the Mixmaster model
Authors:
Riccardo Moriconi,
Giovanni Montani,
Salvatore Capozziello
Abstract:
We study the asymptotic dynamics of the Mixmaster Universe, near the cosmological singularity, considering $f(R)$ gravity up to a quadratic corrections in the Ricci scalar $R$. The analysis is performed in the scalar-tensor framework and adopting Misner-Chitré-like variables to describe the Mixmaster Universe, whose dynamics resembles asymptotically a billiard-ball in a given domain of the half-Po…
▽ More
We study the asymptotic dynamics of the Mixmaster Universe, near the cosmological singularity, considering $f(R)$ gravity up to a quadratic corrections in the Ricci scalar $R$. The analysis is performed in the scalar-tensor framework and adopting Misner-Chitré-like variables to describe the Mixmaster Universe, whose dynamics resembles asymptotically a billiard-ball in a given domain of the half-Poincaré space. The form of the potential well depends on the spatial curvature of the model and on the particular form of the self-interacting scalar field potential. We demonstrate that the potential walls determine an open domain in the configuration region, allowing the point-Universe to reach the absolute of the considered Lobachevsky space. In other words, we outline the existence of a stable final Kasner regime in the Mixmaster evolution, implying the chaos removal near the cosmological singularity. The relevance of the present issue relies both on the general nature of the considered dynamics, allowing its direct extension to the BKL conjecture too, as well as the possibility to regard the considered modified theory of gravity as the first correction to the Einstein-Hilbert action as a Taylor expansion of a generic function $f(R)$ (as soon as a cut-off on the space-time curvature takes place).
△ Less
Submitted 13 November, 2014; v1 submitted 3 November, 2014;
originally announced November 2014.
-
Semiclassical and quantum behavior of the Mixmaster model in the polymer approach
Authors:
Orchidea Maria Lecian,
Giovanni Montani,
Riccardo Moriconi
Abstract:
We analyze the quantum dynamics of the Bianchi Type IX model, as described in the so-called polymer representation of quantum mechanics, to characterize the modifications that a discrete na- ture in the anisotropy variables of the Universe induces on the morphology of the cosmological sin- gularity. We first perform a semiclassical analysis, to be regarded as the zeroth-order approximation of a WK…
▽ More
We analyze the quantum dynamics of the Bianchi Type IX model, as described in the so-called polymer representation of quantum mechanics, to characterize the modifications that a discrete na- ture in the anisotropy variables of the Universe induces on the morphology of the cosmological sin- gularity. We first perform a semiclassical analysis, to be regarded as the zeroth-order approximation of a WKB (Wentzel-Kramers-Brillouin) approximation of the quantum dynamics, and demonstrate how the features of polymer quantum mechanics are able to remove the chaotic properties of the Bianchi IX dynamics. Then, we address the study of the full quantum dynamics of this model in the polymer representation and analyze the two cases, in which the Bianchi IX spatial curvature does not affect the wave-packet behavior, as well as the instance, for which it plays the role of an infinite potential confining the dynamics of the anisotropic variables. The main development of this analysis consists of investigating how, differently from the standard canonical quantum evolution, the high quantum number states are not preserved arbitrarily close to the cosmological singularity. This property emerges as a consequence, on one hand, of the no longer chaotic features of the classical dynamics (on which the Misner analysis is grounded), and, on the other hand, of the impossibility to remove the quantum effect due to the spatial curvature. In the polymer picture, the quantum evolution of the Bianchi IX model remains always significantly far from the semiclassical behavior, as far as both the wave-packet spread and the occupation quantum numbers are concerned. As a result, from a quantum point of view, the Mixmaster dynamics loses any predictivity characterization for the discrete nature of the Universe anisotropy.
△ Less
Submitted 23 November, 2013;
originally announced November 2013.
