检索结果-内蒙古大学图书馆

Improving Effectiveness of Simulation-Based Inference in the Massively Parallel Regime

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 2023年第4期34卷 1100-1114页

作者： Kulkarni, Sourabh Moritz, Csaba Andras Univ Massachusetts Dept Elect & Comp Engn Amherst MA 01003 USA

Simulation-based Inference (SBI) is a widely used set of algorithms to learn the parameters of complex scientific simulation models. While primarily run on CPUs in High-Performance Compute clusters, these algorithms have been shown to scale in performance when developed to be run on massively parallel architectures such as GPUs. While parallelizing existing SBI algorithms provides us with performance gains, this might not be the most efficient way to utilize the achieved parallelism. This work proposes a new parallelism-aware adaptation of an existing SBI method, namely approximate Bayesian computation with Sequential Monte Carlo(ABC-SMC). This new adaptation is designed to utilize the parallelism not only for performance gain, but also toward qualitative benefits in the learnt parameters. The key idea is to replace the notion of a single 'step-size' hyperparameter, which governs how the state space of parameters is explored during learning, with step-sizes sampled from a tuned Beta distribution. This allows this new ABC-SMC algorithm to more efficiently explore the state-space of the parameters being learned. We test the effectiveness of the proposed algorithm to learn parameters for an epidemiology model running on a Tesla T4 GPU. Compared to the parallelized state-of-the-art SBI algorithm, we get similar quality results in similar to 100x fewer simulations and observe similar to 80x lower run-to-run variance across 10 independent trials.

关键词： Inference algorithms Computational modeling Biological system modeling approximation algorithms Perturbation methods Optimization Bayes methods Approximate Bayesian computation Bayesian inference COVID-19 compartmental models epidemiology GPU high performance computing likelihood-free inference parallel algorithms simulation-based inference statistical machine learning stochastic optimization

来源：评论

学校读者我要写书评

暂无评论

Recursive algorithms for the Estimation of Multiple Superimposed Undamped Tones and Their Application to Radar Systems

引用

IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS 2023年第2期59卷 1834-1853页

作者： Di Viesti, Pasquale Davoli, Alessandro Guerzoni, Giorgio Vitetta, Giorgio M. M. Univ Modena & Reggio Emilia Dept Engn Enzo Ferrari I-41125 Modena Italy Consorzio Nazl Interuniv Perle Telecomunicaz I-43124 Parma Italy

In this article, two recursive algorithms for the detection of multiple superimposed tones in noise and the estimation of their parameters are derived. They are based on a maximum likelihood approach and combine an innovative single-tone estimator with a serial cancellation procedure. Our numerical results lead to the conclusion that the developed methods can achieve a substantially better accuracy-complexity tradeoff than various related techniques in the presence of multiple closely spaced tones. Moreover, they can be exploited to detect multiple closely spaced targets and estimate their spatial coordinates in multiple-input multiple-output frequency-modulated continuous wave radar systems.

关键词： Frequency estimation Maximum likelihood estimation Radar Iterative methods approximation algorithms MIMO communication Interpolation maximum likelihood (ML) estimation multiple-input multiple-output (MIMO) systems radar spectral analysis

来源：评论

学校读者我要写书评

暂无评论

An ϵ-approximation On-line Algorithm for Quantiles Based on Counters 14

An ϵ-Approximation On-line Algorithm for Quantiles Based on...

引用

14th IIAI International Congress on Advanced Applied Informatics, IIAI-AAI 2023

作者： Maeda, Kosuke Iwanuma, Koji University of Yamanashi Kofu-shi Japan

ISBN: (纸本)9798350324228

The quantiles are very important as a statistic that gives an overview of big data. In this paper, we study the GK algorithm for finding ϵ-approximate quantiles and point out its redundancy hidden the quantile summary. We introduce a counter mechanism and give a new approximation algorithm that is more space efficient. We theoretically guarantee that the maximum quantile error is within the range of the ϵ-approximation, and confirm by computational experiments that the spatial complexity required to preserve the ϵ-approximate quantile summary is significantly improved. © 2023 IEEE.

关键词： approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

Fixed-Parameter algorithms for Unsplittable Flow Cover

引用

THEORY OF COMPUTING SYSTEMS 2023年第1期67卷 89-124页

作者： Cristi, Andres Mari, Mathieu Wiese, Andreas Univ Chile Santiago Chile Univ PSL Ecole Normale Super Paris France

The Unsplittable Flow Cover problem (UFP-cover) models the well-studied general caching problem and various natural resource allocation settings. We are given a path with a demand on each edge and a set of tasks, each task being defined by a subpath and a size. The goal is to select a subset of the tasks of minimum cardinality such that on each edge e the total size of the selected tasks using e is at least the demand of e. There is a polynomial time 4-approximation for the problem (Bar-Noy et al. STOC 2001) and also a QPTAS (Hohn et al. ICALP 2018). In this paper we study fixed-parameter algorithms for the problem. We show that it is W[1]-hard but it becomes FPT if we can slighly violate the edge demands (resource augmentation) and also if there are at most k different task sizes. Then we present a parameterized approximation scheme (PAS), i.e., an algorithm with a running time of f (k) center dot n(O epsilon)(1) that outputs a solution with at most (1 + tasks or asserts that there is no solution with at most k tasks. In this algorithm we use a new trick that intuitively allows us to pretend that we can select tasks from OPT multiple times. We show that the other two algorithms extend also to the weighted case of the problem, at the expense of losing a factor of 1 + in the cost of the selected tasks.

关键词： Unsplittable flow cover Fixed parameter algorithms approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

An Efficient Impulsive Adaptive Dynamic Programming Algorithm for Stochastic Systems

引用

IEEE TRANSACTIONS ON CYBERNETICS 2023年第9期53卷 5545-5559页

作者： Liang, Mingming Wang, Yonghua Liu, Derong Guangdong Univ Technol Sch Automat Guangzhou 510006 Peoples R China Univ Illinois Dept Elect & Comp Engn Chicago IL 60607 USA

In this study, a novel general impulsive transition matrix is defined, which can reveal the transition dynamics and probability distribution evolution patterns for all system states between two impulsive ``events,'' instead of two regular time indexes. Based on this general matrix, the policy iteration-based impulsive adaptive dynamic programming (IADP) algorithm along with its variant, which is a more efficient IADP (EIADP) algorithm, are developed in order to solve the optimal impulsive control problems of discrete stochastic systems. Through analyzing the monotonicity, stability, and convergency properties of the obtained iterative value functions and control laws, it is proved that the IADP and EIADP algorithms both converge to the optimal impulsive performance index function. By dividing the whole impulsive policy into smaller pieces, the proposed EIADP algorithm updates the iterative policies in a ``piece-by-piece'' manner according to the actual hardware constraints. This feature of the EIADP method enables these ADP-based algorithms to be fully optimized to run on all ``sizes'' of computing devices including the ones with low memory spaces. A simulation experiment is conducted to validate the effectiveness of the present methods.

关键词： Heuristic algorithms Stochastic systems approximation algorithms Aerospace electronics Markov processes Dynamic programming Probability distribution Adaptive dynamic programming (ADP) impulsive stochastic systems optimal control policy iteration transition matrix

来源：评论

学校读者我要写书评

暂无评论

Distributed Bayesian Inference Over Sensor Networks

引用

IEEE TRANSACTIONS ON CYBERNETICS 2023年第3期53卷 1587-1597页

作者： Ye, Baijia Qin, Jiahu Fu, Weiming Zhu, Yingda Wang, Yaonan Kang, Yu Univ Sci & Technol China Dept Automat Hefei 230027 Peoples R China Hefei Comprehens Natl Sci Ctr Inst Artificial Intelligence Hefei 230088 Peoples R China Baidu Mobile Ecosyst Grp Shanghai 201210 Peoples R China Hunan Univ Coll Elect & Informat Engn Changsha 410082 Hunan Peoples R China Natl Engn Lab Robot Visual Percept & Control Tech Changsha 410082 Hunan Peoples R China

In this article, two novel distributed variational Bayesian (VB) algorithms for a general class of conjugate-exponential models are proposed over synchronous and asynchronous sensor networks. First, we design a penalty-based distributed VB (PB-DVB) algorithm for synchronous networks, where a penalty function based on the Kullback-Leibler (KL) divergence is introduced to penalize the difference of posterior distributions between nodes. Then, a token-passing-based distributed VB (TPB-DVB) algorithm is developed for asynchronous networks by borrowing the token-passing approach and the stochastic variational inference. Finally, applications of the proposed algorithm on the Gaussian mixture model (GMM) are exhibited. Simulation results show that the PB-DVB algorithm has good performance in the aspects of estimation/inference ability, robustness against initialization, and convergence speed, and the TPB-DVB algorithm is superior to existing token-passing-based distributed clustering algorithms.

关键词： Inference algorithms Clustering algorithms Convergence Bayes methods Hidden Markov models Biological system modeling approximation algorithms Conjugate-exponential model distributed variational Bayesian (VB) Kullback-Leibler (KL) divergence sensor networks token-passing method

来源：评论

学校读者我要写书评

暂无评论

Cost-Aware and Distance-Constrained Collective Spatial Keyword Query

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING 2023年第2期35卷 1324-1336页

作者： Chan, Harry Kai-Ho Liu, Shengxin Long, Cheng Wong, Raymond Chi-Wing Roskilde Univ Dept People & Technol Roskilde 4000 Denmark Harbin Inst Technol Sch Comp Sci & Technol Shenzhen 518055 Peoples R China Nanyang Technol Univ Sch Comp Sci & Engn Singapore 639798 Singapore Hong Kong Univ Sci & Technol Dept Comp Sci & Engn Kowloon Hong Kong Peoples R China

With the proliferation of location-based services, geo-textual data is becoming ubiquitous. Objects involved in geo-textual data include geospatial locations, textual descriptions or keywords, and various attributes (e.g., a point-of-interest has its expenses and users' ratings). Many types of spatial keyword queries have been proposed on geo-textual data. Among them, one prominent type is to find, for a query consisting of a query location and some query keywords, a set of multiple objects such that the objects in the set collectively cover all the query keywords and the object set is of good quality according to some criteria. Existing studies define the criteria either based on the geospatial information of the objects solely or simply treat the geospatial information and the attribute information of the objects together without differentiation though they may have different semantics and scales. As a result, they cannot provide users flexibility to express finer grained preferences on the objects. In this paper, we propose a new criterion which is to find a set of objects where the distance (defined based on the geospatial information) is at most a threshold specified by users and the cost (defined based on the attribute information) is optimized. We develop a suite of two algorithms including an exact algorithm and an approximation algorithm with provable guarantees for the problem. We conducted extensive experiments on real datasets which verified the efficiency and effectiveness of proposed algorithms.

关键词： Geospatial analysis approximation algorithms Semantics Cost function Spatial databases Search problems Mice Spatial keyword queries spatial database query processing

来源：评论

学校读者我要写书评

暂无评论

Robust Multi-Robot Active Target Tracking Against Sensing and Communication Attacks

引用

IEEE TRANSACTIONS ON ROBOTICS 2023年第3期39卷 1768-1780页

作者： Zhou, Lifeng Kumar, Vijay Univ Penn GRASP Lab Philadelphia PA 19104 USA Drexel Univ Dept Elect & Comp Engn Philadelphia PA 19104 USA

The problem of multi-robot target tracking asks for actively planning the joint motion of robots to track targets. In this article, we focus on such target tracking problems in adversarial environments, where attacks or failures may deactivate robots' sensors and communications. In contrast to the previous works that consider no attacks or sensing attacks only, we formalize the first robust multi-robot tracking framework that accounts for any fixed numbers of worst-case sensing and communication attacks. To secure against such attacks, we design the first robust planning algorithm, named Robust Active Target Tracking (RATT), which approximates the communication attacks to equivalent sensing attacks and then optimizes against the approximated and original sensing attacks. We show that RATT provides provable suboptimality bounds on the tracking quality for any non-decreasing objective function. Our analysis utilizes the notations of curvature for set functions introduced in combinatorial optimization. In addition, RATT runs in polynomial time and terminates with the same running time as state-of-the-art algorithms for (non-robust) target tracking. Finally, we evaluate RATT with both the qualitative and quantitative simulations across various scenarios. In the evaluations, RATT exhibits a tracking quality that is near-optimal and superior to varying non-robust heuristics. We also demonstrate RATT's superiority and robustness against varying attack models (e.g., worst-case and bounded rational attacks) and with over- and under-estimated numbers of attacks.

关键词： Robot sensing systems Robots Sensors Target tracking Robot kinematics Noise measurement approximation algorithms Active target tracking algorithm design and analysis combinatorial optimization multi-robot systems robotics in adversarial environments

来源：评论

学校读者我要写书评

暂无评论

Event-Triggered Prescribed Performance Fuzzy Fault-Tolerant Control for Unknown Euler-Lagrange Systems With Any Bounded Initial Values

引用

IEEE TRANSACTIONS ON FUZZY SYSTEMS 2023年第6期31卷 2065-2075页

作者： Hu, Yunsong Yan, Huaicheng Zhang, Youmin Zhang, Hao Chang, Yufang East China Univ Sci & Technol Minist Educ Key Lab Smart Mfg Energy Chem Proc Shanghai 200237 Peoples R China Hubei Univ Technol Hubei Key Lab High Efficiency Utilizat Solar Energ Wuhan 430068 Peoples R China Concordia Univ Dept Mech Ind & Aerosp Engn Montreal PQ H3G 1M8 Canada Tongji Univ Dept Control Sci & Engn Shanghai 200092 Peoples R China

This article investigates the tracking problem of event-triggered prescribed performance fuzzy fault-tolerant control (FTC) for unknown Euler-Lagrange systems with actuator faults and external disturbances. First, the barrier Lyapunov functions (BLFs) and prescribed performance functions are synthesized to guarantee that the tracking errors satisfy the preset transient performance. Different from existing prescribed performance control methods, which require the initial values of the tracking errors to be within the prescribed performance functions, an error transformation method is introduced to ensure that the tracking errors with any bounded initial values can enter the preset boundaries within a preset time. Then, considering the unavailability of system parameters, the fuzzy logic systems are used to approximate unknown parameters of the system. What is more, to solve the problem of limited communication and computing resources in practical systems, an improved event-triggered control (ETC) scheme is proposed, which can reduce the communication and computation burden without satisfying the input-to-state stability assumption. Meanwhile, the Zeno phenomenon can be avoided. Furthermore, the effects of actuator faults and the event-triggered mechanism are handled by Nussbaum gain technology. Finally, the superiority of the proposed control algorithm is verified by simulation results.

关键词： Actuators Nonlinear systems Fault tolerant systems Fault tolerance approximation algorithms Transient analysis Power system stability Barrier Lyapunov function (BLF) event-triggered control (ETC) fault-tolerant control (FTC) fuzzy logic systems (FLSs) prescribed performance

来源：评论

学校读者我要写书评

暂无评论

Placement and Allocation of Virtual Network Functions: Multi-Dimensional Case

引用

IEEE TRANSACTIONS ON MOBILE COMPUTING 2023年第7期22卷 4195-4211页

作者： Sallam, Gamal Zheng, Zizhan Ji, Bo Temple Univ Dept Comp & Informat Sci Philadelphia PA 19122 USA Tulane Univ Dept Comp Sci New Orleans LA 70118 USA Virginia Tech Dept Comp Sci Blacksburg VA 24061 USA

Network function virtualization (NFV) is an emerging design paradigm that replaces physical middlebox devices with software modules running on general purpose commodity servers. While gradually transitioning to NFV, Internet service providers face the problem of where to introduce NFV in order to make the most benefit of that;here, we measure the benefit by the amount of traffic that can be served in an NFV-enabled network. This problem is non-trivial as it is composed of two challenging subproblems: 1) placement of nodes to support virtual network functions (referred to as VNF-nodes);2) allocation of the VNF-nodes' resources to network flows. These two subproblems must be jointly considered to satisfy the objective of serving the maximum amount of traffic. This problem has been studied for the one-dimensional setting, where all network flows require one network function, which requires a unit of resource to process a unit of flow. In this work, we consider the multi-dimensional setting, where flows must be processed by multiple network functions, which require a different amount of each resource to process a unit of flow. The multi-dimensional setting introduces new challenges in addition to those of the one-dimensional setting (e.g., NP-hardness and non-submodularity) and also makes the resource allocation subproblem a multi-dimensional generalization of the generalized assignment problem with assignment restrictions. To address these difficulties, we propose a novel two-level relaxation method that allows us to draw a connection to the sequence submodular theory and utilize the property of sequence submodularity along with the primal-dual technique to design two approximation algorithms. We further prove that the proposed algorithms have a non-trivial approximation ratio that depends on the number of VNF-nodes, resources, and a measure of the available resource compared to flow demand. Finally, we perform trace-driven simulations to show the effectiveness of the

关键词： Virtual network functions multi-dimensional resource allocation sequence submodular functions approximation algorithms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：