检索结果-内蒙古大学图书馆

Hierarchic neighbors embedding

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Liu, Shenglan Yu, Yang Liu, Yang Qiao, Hong Feng, Lin Feng, Jiashi School of Computer Science and Technology Faculty of Electronic Information and Electrical Engineering Dalian University of Technology Dalian Liaoning116024 China State Key Laboratory of Management and Control for Complex Systems Institute of Automation Chinese Academy of Sciences Beijing100190 China Department of Electrical and Computer Engineering National University of Singapore Singapore Singapore

Manifold learning now plays a very important role in machine learning and many relevant applications. Although its superior performance in dealing with nonlinear data distribution, data sparsity is always a thorny knot. There are few researches to well handle it in manifold learning. In this paper, we propose Hierarchic Neighbors Embedding (HNE), which enhance local connection by the hierarchic combination of neighbors. After further analyzing topological connection and reconstruction performance, three different versions of HNE are given. The experimental results show that our methods work well on both synthetic data and high-dimensional real-world tasks. HNE develops the outstanding advantages in dealing with general data. Furthermore, comparing with other popular manifold learning methods, the performance on sparse samples and weak-connected manifolds is better for HNE. Copyright © 2019, The Authors. All rights reserved.

关键词： Topology

Local Exchangeability

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Campbell, Trevor Syed, Saifuddin Yang, Chiao-Yu Jordan, Michael I. Broderick, Tamara Department of Statistics University of British Columbia Vancouver Canada Department of Electrical Engineering and Computer Science University of California Berkeley Berkeley United States Laboratory for Information and Decision Systems Massachusetts Institute of Technology Cambridge United States

Exchangeability—in which the distribution of an infinite sequence is invariant to reorderings of its elements—implies the existence of a simple conditional independence structure that may be leveraged in the design of statistical models and inference procedures. In this work, we study a relaxation of exchangeability in which this invariance need not hold precisely. We introduce the notion of local exchangeability—where swapping data associated with nearby covariates causes a bounded change in the distribution. We prove that locally exchangeable processes correspond to independent observations from an underlying measure-valued stochastic process. Using this main probabilistic result, we show that the local empirical measure of a finite collection of observations provides an approximation of the underlying measure-valued process and Bayesian posterior predictive distributions. The paper concludes with applications of the main theoretical results to a model from Bayesian nonparametrics and covariate-dependent permutation tests. Copyright © 2019, The Authors. All rights reserved.

关键词： Random processes

Multi-junction laser power converters exceeding 50% efficiency in the short wavelength infrared

学校读者我要写书评

暂无评论

Cell Reports Physical Science 2025年

作者： Gavin P. Forcade D. Paige Wilson Meghan N. Beattie Carmine Pellegrino Henning Helmers Robert F.H. Hunter Oliver Hӧhn David Lackner Louis-Philippe St-Arnaud Thomas N.D. Tibbits Daniel Poitras Christopher E. Valdivia Yuri Grinberg Alexandre W. Walker Jacob J. Krich Karin Hinzer Department of Physics University of Ottawa Ottawa ON K1N 7P9 Canada SUNLAB School of Electrical Engineering and Computer Sciences University of Ottawa Ottawa ON K1N 7P9 Canada Fraunhofer Institute for Solar Energy Systems ISE 79110 Freiburg Germany Quantum and Nanotechnologies Research Center National Research Council of Canada Ottawa ON K1K 4P7 Canada Digital Technologies Research Center National Research Council of Canada Ottawa ON K1K 4P7 Canada

Photonic or laser power converters are crucial components in power-by-light systems. However, their use in long-distance applications has been hindered by low efficiencies and output voltages within the optical fiber transmission window of 1.3–1.6 μm laser wavelengths. Here, we improve and simplify the design and characterization processes for photonic power converters, exceeding 50% efficiency under 1.446 μm laser light. We develop a calibrated model predicting efficiency gains with increasing bandgap, reaching up to 57% efficiency at a 1.3-μm wavelength. As a first demonstration, we produce a high-efficiency device designed by the model: a four-junction InGaAsP photonic power converter with a conversion efficiency of 53.6% ± 1.3% and an output voltage above 2 V under 15.2 W/cm 2 of 1.446 μm laser light. These advances open new, practical pathways for integrating photonic power converters into telecommunication systems and unlock the potential to further optimize their design with machine learning algorithms trained with our predictive model.

关键词： laser power converter photonic power converter optical power converter III-V photovoltaics InGaAs InGaAsP InP multi-junction optoelectronics telecommunications power transmission photonics

Machine learning prediction of critical transition and system collapse

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Kong, Ling-Wei Fan, Hua-Wei Grebogi, Celso Lai, Ying-Cheng School of Electrical Computer and Energy Engineering Arizona State University TempeAZ85287 United States School of Physics and Information Technology Shaanxi Normal University Xi’an710062 China Institute for Complex Systems and Mathematical Biology School of Natural and Computing Sciences King’s College University of Aberdeen United Kingdom Department of Physics Arizona State University TempeAZ85287 United States

To predict a critical transition due to parameter drift without relying on model is an outstanding problem in nonlinear dynamics and applied fields. A closely related problem is to predict whether the system is already in or if the system will be in a transient state preceding its collapse. We develop a model free, machine learning based solution to both problems by exploiting reservoir computing to incorporate a parameter input channel. We demonstrate that, when the machine is trained in the normal functioning regime with a chaotic attractor (i.e., before the critical transition), the transition point can be predicted accurately. Remarkably, for a parameter drift through the critical point, the machine with the input parameter channel is able to predict not only that the system will be in a transient state, but also the average transient time before the final collapsea. Copyright © 2020, The Authors. All rights reserved.

关键词： Forecasting

Motion Planning In High-Dimensional Spaces

学校读者我要写书评

暂无评论

arXiv 2018年

作者： Petrović, Luka University of Zagreb Faculty of Electrical Engineering and Computing Department of Control and Computer Engineering Laboratory for Autonomous Systems Mobile Robotics

—Motion planning is a key tool that allows robots to navigate through an environment without collisions. The problem of robot motion planning has been studied in great detail over the last several decades, with researchers initially focusing on systems such as planar mobile robots and low degree-of-freedom (DOF) robotic arms. The increased use of high DOF robots that must perform tasks in real time in complex dynamic environments spurs the need for fast motion planning algorithms. In this overview, we discuss several types of strategies for motion planning in high dimensional spaces and dissect some of them, namely grid search based, sampling based and trajectory optimization based approaches. We compare them and outline their advantages and disadvantages, and finally, provide an insight into future research opportunities. Copyright © 2018, The Authors. All rights reserved.

关键词： Motion planning

Self-Calibration for massive MIMO with channel reciprocity and channel estimation errors

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Mi, De Zhang, Lei Dianati, Mehrdad Muhaidat, Sami Xiao, Pei Tafazolli, Rahim Institute for Communication Systems University of Surrey United Kingdom School of Engineering University of Glasgow United Kingdom Warwick Manufacturing Group University of Warwick United Kingdom Department of Electrical and Computer Engineering Khalifa University United Arab Emirates

—In time-division-duplexing (TDD) massive multiple-input multiple-output (MIMO) systems, channel reciprocity is exploited to overcome the overwhelming pilot training and the feedback overhead. However, in practical scenarios, the imperfections in channel reciprocity, mainly caused by radio-frequency mismatches among the antennas at the base station side, can significantly degrade the system performance and might become a performance limiting factor. In order to compensate for these imperfections, we present and investigate two new calibration schemes for TDD-based massive multi-user MIMO systems, namely, relative calibration and inverse calibration. In particular, the design of the proposed inverse calibration takes into account a compound effect of channel reciprocity error and channel estimation error. We further derive closed-form expressions for the ergodic sum rate, assuming maximum ratio transmissions with the compound effect of both errors. We demonstrate that the inverse calibration scheme outperforms the traditional relative calibration scheme. The proposed analytical results are also verified by simulated illustrations. Copyright © 2019, The Authors. All rights reserved.

关键词： Channel estimation

A fuzzy inference system for disturbance classification in power distribution networks using a low-cost PMU

学校读者我要写书评

暂无评论

A fuzzy inference system for disturbance classification in p...

2018 IEEE International Conference on Fuzzy systems, FUZZ 2018

作者： Sobrinho, André Sanches Fonseca Flauzino, Rogério Andrade Liboni, Luisa H.B. Maques Costa, Eduardo C. Spatti, Danilo Hernane Department of Electrical Engineering Federal University of Technology Paraná Cornélio Procópio Brazil Department of Electrical and Computer Engineering University of São Paulo São Carlos Brazil Department of Electrical and Computer Engineering Federal Institute of Education Science and Technology of São Paulo Sertãozinho Brazil Department of Energy and Automation Engineering University of São Paulo São Paulo Brazil Department of Computer Systems University of São Paulo São Carlos Brazil

ISBN: (纸本)9781509060207

electrical distribution networks are evolving at a high pace, mainly due to the increase of distributed energy sources, the growing use of advanced metering infrastructures and new management systems. Phasor Measurement Units (PMUs) have the potential to play an essential role in the monitoring of power distribution systems. This paper proposes a fuzzy inference system to be embedded in a low-cost PMU for classifying the origin/cause of electrical disturbances in power distribution systems. Precise knowledge of the origin of electrical disturbances is fundamental to meet the requirements of load and network modeling tasks. Experimental results show the effective operation of the fuzzy inference system and its ability to address the measurement inaccuracies of the PMU. Furthermore, a remote server was used to correctly estimate electrical parameters of loads by using the output information of the fuzzy inference system. © 2018 IEEE.

关键词： Fuzzy systems

Controlling symmetries and clustered dynamics of complex networks

学校读者我要写书评

暂无评论

arXiv 2020年

作者： Gambuzza, Lucia Valentina Frasca, M. Sorrentino, F. Pecora, L.M. Boccaletti, S. Dipartimento di Ingegneria Elettrica Elettronica e Informatica University of Catania Catania95029 Italy Dipartimento di Ingegneria Elettrica Elettronica e Informatica University of Catania Catania95029 Italy CNR-IASI Institute for Systems Analysis and Computer Science A. Ruberti Rome00185 Italy Department of Mechanical Engineering Univ. of New Mexico AlbuquerqueNM87131 United States Dep. of Electrical and Computer Engineering Univ. of New Mexico AlbuquerqueNM87131 United States U.S. Naval Research Laboratory WashingtonDC20375 United States CNR Institute of Complex Systems Via Madonna del Piano 10 Florence50019 Italy Unmanned Systems Research Institute Northwestern Polytechnical University Xi’an710072 China Moscow Institute of Physics and Technology Dolgoprudny Moscow Region141701 Russia

Symmetries are an essential feature of complex networks as they regulate how the graph collective dynamics organizes into clustered states. We here show how to control network symmetries, and how to enforce patterned states of synchronization with nodes clustered in a desired way. Our approach consists of perturbing the original network connectivity, either by adding new edges or by adding/removing links together with modifying their weights. By solving suitable optimization problems, we furthermore guarantee that changes made on the existing topology are minimal. The conditions for the stability of the enforced pattern are derived for the general case, and the performance of the method is illustrated with paradigmatic examples. Our results are relevant to all the practical situations in which coordination of the networked systems into diverse groups may be desirable, such as for teams of robots, unmanned autonomous vehicles, power grids and central pattern generators. Copyright © 2020, The Authors. All rights reserved.

关键词： Complex networks

Black box submodular maximization: Discrete and continuous settings

学校读者我要写书评

暂无评论

arXiv 2019年

作者： Chen, Lin Zhang, Mingrui Hassani, Hamed Karbasi, Amin Department of Electrical Engineering Yale Institute for Network Science Yale University New HavenCT06520 United States Department of Statistics and Data Science Yale University New HavenCT06511 United States Department of Electrical and Systems Engineering University of Pennsylvania PhiladelphiaPA19104 United States Department of Electrical Engineering and Computer Science Yale University New HavenCT06520 United States

In this paper, we consider the problem of black box continuous submodular maximization where we only have access to the function values and no information about the derivatives is provided. For a monotone and continuous DR-submodular function, and subject to a bounded convex body constraint, we propose Black-box Continuous Greedy, a derivative-free algorithm that provably achieves the tight [(1 − 1/e)OPT − Ε] approximation guarantee with O(d/Ε3) function evaluations. We then extend our result to the stochastic setting where function values are subject to stochastic zero-mean noise. It is through this stochastic generalization that we revisit the discrete submodular maximization problem and use the multi-linear extension as a bridge between discrete and continuous settings. Finally, we extensively evaluate the performance of our algorithm on continuous and discrete submodular objective functions using both synthetic and real data. Copyright © 2019, The Authors. All rights reserved.

关键词： Function evaluation