Filters

2021 - 20224

More filters

2022 - 20234
Reset filters

Settings

search.filters.applied.f.access: openAccess × End date: 2024 × Start date: 2020 × Version: publishedVersion × Publisher: Amsterdam : Elsevier × WGL Institute: WIAS ×

Search Results

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item
    Revealing all states of dewetting of a thin gold layer on a silicon surface by nanosecond laser conditioning
    (Amsterdam : Elsevier, 2021) Ernst, Owen C.; Uebel, David; Kayser, Stefan; Lange, Felix; Teubner, Thomas; Boeck, Torsten
    Dewetting is a ubiquitous phenomenon which can be applied to the laser synthesis of nanoparticles. A classical spinodal dewetting process takes place in four successive states, which differ from each other in their morphology. In this study all states are revealed by interaction of pulsed nanosecond UV laser light with thin gold layers with thicknesses between 1 nm and 10 nm on (100) silicon wafers. The specific morphologies of the dewetting states are discussed with particular emphasis on the state boundaries. The main parameter determining which state is formed is not the duration for which the gold remains liquid, but rather the input energy provided by the laser. This shows that each state transition has a separate measurable activation energy. The temperature during the nanosecond pulses and the duration during which the gold remains liquid was determined by simulation using the COMSOL Multiphysics® software package. Using these calculations, an accurate local temperature profile and its development over time was simulated. An analytical study of the morphologies and formed structures was performed using Minkowski measures. With aid of this tool, the laser induced structures were compared with thermally annealed samples, with perfectly ordered structures and with perfectly random structures. The results show that both, structures of the laser induced and the annealed samples, strongly resemble the perfectly ordered structures. This reveals a close relationship between these structures and suggests that the phenomenon under investigation is indeed a spinodal dewetting generated by an internal material wave function. The purposeful generation of these structures and the elucidation of the underlying mechanism of dewetting by short pulse lasers may assist the realisation of various technical elements such as nanowires in science and industry. © 2020
  • Thumbnail Image
    Item
    Hyperfast second-order local solvers for efficient statistically preconditioned distributed optimization
    (Amsterdam : Elsevier, 2022) Dvurechensky, Pavel; Kamzolov, Dmitry; Lukashevich, Aleksandr; Lee, Soomin; Ordentlich, Erik; Uribe, César A.; Gasnikov, Alexander
    Statistical preconditioning enables fast methods for distributed large-scale empirical risk minimization problems. In this approach, multiple worker nodes compute gradients in parallel, which are then used by the central node to update the parameter by solving an auxiliary (preconditioned) smaller-scale optimization problem. The recently proposed Statistically Preconditioned Accelerated Gradient (SPAG) method [1] has complexity bounds superior to other such algorithms but requires an exact solution for computationally intensive auxiliary optimization problems at every iteration. In this paper, we propose an Inexact SPAG (InSPAG) and explicitly characterize the accuracy by which the corresponding auxiliary subproblem needs to be solved to guarantee the same convergence rate as the exact method. We build our results by first developing an inexact adaptive accelerated Bregman proximal gradient method for general optimization problems under relative smoothness and strong convexity assumptions, which may be of independent interest. Moreover, we explore the properties of the auxiliary problem in the InSPAG algorithm assuming Lipschitz third-order derivatives and strong convexity. For such problem class, we develop a linearly convergent Hyperfast second-order method and estimate the total complexity of the InSPAG method with hyperfast auxiliary problem solver. Finally, we illustrate the proposed method's practical efficiency by performing large-scale numerical experiments on logistic regression models. To the best of our knowledge, these are the first empirical results on implementing high-order methods on large-scale problems, as we work with data where the dimension is of the order of 3 million, and the number of samples is 700 million.
  • Thumbnail Image
    Item
    First-Order Methods for Convex Optimization
    (Amsterdam : Elsevier, 2021) Dvurechensky, Pavel; Shtern, Shimrit; Staudigl, Mathias
    First-order methods for solving convex optimization problems have been at the forefront of mathematical optimization in the last 20 years. The rapid development of this important class of algorithms is motivated by the success stories reported in various applications, including most importantly machine learning, signal processing, imaging and control theory. First-order methods have the potential to provide low accuracy solutions at low computational complexity which makes them an attractive set of tools in large-scale optimization problems. In this survey, we cover a number of key developments in gradient-based optimization methods. This includes non-Euclidean extensions of the classical proximal gradient method, and its accelerated versions. Additionally we survey recent developments within the class of projection-free methods, and proximal versions of primal-dual schemes. We give complete proofs for various key results, and highlight the unifying aspects of several optimization algorithms.
  • Thumbnail Image
    Item
    Accelerated variance-reduced methods for saddle-point problems
    (Amsterdam : Elsevier, 2022) Borodich, Ekaterina; Tominin, Vladislav; Tominin, Yaroslav; Kovalev, Dmitry; Gasnikov, Alexander; Dvurechensky, Pavel
    We consider composite minimax optimization problems where the goal is to find a saddle-point of a large sum of non-bilinear objective functions augmented by simple composite regularizers for the primal and dual variables. For such problems, under the average-smoothness assumption, we propose accelerated stochastic variance-reduced algorithms with optimal up to logarithmic factors complexity bounds. In particular, we consider strongly-convex-strongly-concave, convex-strongly-concave, and convex-concave objectives. To the best of our knowledge, these are the first nearly-optimal algorithms for this setting.