Search results for " convergence."

Interpretable Option Discovery Using Deep Q-Learning and Variational Autoencoders

2021

Deep Reinforcement Learning (RL) is unquestionably a robust framework to train autonomous agents in a wide variety of disciplines. However, traditional deep and shallow model-free RL algorithms suffer from low sample efficiency and inadequate generalization for sparse state spaces. The options framework with temporal abstractions [18] is perhaps the most promising method to solve these problems, but it still has noticeable shortcomings. It only guarantees local convergence, and it is challenging to automate initiation and termination conditions, which in practice are commonly hand-crafted.

Institutions and Geography: Empirical tests of Spatial Growth Models for the regions of Europe

2010

This article provides an empirical assessment of the growth experiences of European regions, during the period 1991–2004, by taking into account the spatial effects due to both institutions and geography. These effects have been modelled by means of specific controls and by using a non-conventional spatial weight matrix. Results favour a model dealing with substantive spatial externalities. Within this framework, the country-specific institutions are strongly and positively related to the regional productivity's growth rate. In addition, the geo-institutional proximity increases the spatial dependence of the regional output per worker and raises the speed of convergence. By contrast, the pu…

Great Recession and club convergence in Europe: A cross‐country, cross‐region panel analysis (2000–2015)

2020

The paper aims at investigating the impact of the Great Recession on per capita GDP convergence process across European regions and countries. Using the time-varying factor model developed by Phillips and Sul for the period 2000–2015 and two different merging procedures to identify clubs, we provide evidence of the diverging impact of the Great Recession “between” the higher and the lower convergence clubs at both regional and country levels as well as of the strengthening of the convergence process “within” most clubs. In addition, we add further evidence to the common belief of a “multi-speed” Europe by contrasting Eastern European countries' and regions' behavior vis-à-vis original Europ…

Fast Image Restoration Algorithms Based on PDE Models Using Modified Hopfield Neural Network

2010

Two image restoration algorithms based on modified Hop field neural network and variational partial differential equations (PDE) were proposed in our previous work [1, 2]. But the convergence rate of the proposed algorithms was slow. In this paper, we develop a fast update rule based on modified Hop field neural network (MHNN) of continuous state change and two fast image restoration algorithms. Experimental results show that, when compared with the previous algorithms, our proposed algorithms have better performance both in convergence rate and in image restoration quality.

Spectral Density Estimate for Stable Processes Observed with an Additive Error

2018

International audience; In this paper, a symmetric alpha stable process where its spectral representation has an additive error is considered. The error is supposed to be constant. A periodogram as estimator of the spectral density and its rate of convergence are given. In order to give an asymptotically unbiased and consistent estimate of the spectral density, this periodogram is smoothed by an adapted spectral window. The rate of convergence is given.

Technological differences and convergence in the OECD

2000

Abstract. In this paper we test the homogeneity of the technological parameters among OECD countries, which is the maintained hypothesis in most of the empirical growth literature. We first identify differences in the constant term of the convergence equation estimated for the OECD 1960/1990 sample using a fixed- effects estimator. Then we provide a formal test of the homogeneity of technological parameters across groups of countries. We identify at least two different groups within the OECD, with significantly different technologies. Convergence within each group is fast, supporting the notion of club convergence. Nevertheless, the implausible parameter values obtained for the leading tech…

Mapping discounted and undiscounted Markov Decision Problems onto Hopfield neural networks

1995

This paper presents a framework for mapping the value-iteration and related successive approximation methods for Markov Decision Problems onto Hopfield neural networks, for both discounted and undiscounted versions of the finite state and action spaces. We analyse the asymptotic behaviour of the control sets and we give some estimates on the convergence rate for the value-iteration scheme. We relate the convergence properties on an energy function which represents the key point in mapping Markov Decision Problems onto Hopfield networks. Finally, an application from queueing systems in communication networks is taken into consideration and the results of computer simulation of Hopfield netwo…

Direct perturbation theory in terms of energy derivatives: scalar-relativistic treatment up to sixth order.

2011

A formulation of sixth-order direct perturbation theory (DPT) to treat relativistic effects in quantum-chemical calculations is presented in the framework of derivative theory. Detailed expressions for DPT6 are given at the Hartree-Fock level in terms of the third derivative of the energy with respect to the relativistic perturbation parameter defined as λ(rel)=c(-2). They were implemented for the computation of scalar-relativistic energy corrections. The convergence of the scalar-relativistic DPT expansion is studied for energies and first-order properties such as dipole moment and electric-field gradient within the series of the hydrogen halides (HX, X = F, Cl, Br, I, and At). Comparison …

Boundary regularity and the uniform convergence of quasiconformal mappings

1979

A new result on impulsive differential equations involving non-absolutely convergent integrals

2009

AbstractIn this paper we obtain, as an application of a Darbo-type theorem, global solutions for differential equations with impulse effects, under the assumption that the function on the right-hand side is integrable in the Henstock sense. We thus generalize several previously given results in literature, for ordinary or impulsive equations.