检索结果-内蒙古大学图书馆

Optimal Sample Acquisition for Optimally Weighted PCA From Heterogeneous Quality Sources

IEEE SIGNAL PROCESSING LETTERS 2025年 32卷 1425-1429页

作者： Hong, David Balzano, Laura Univ Delaware Dept Elect & Comp Engn Newark DE 19716 USA Univ Michigan Dept Elect Engn & Comp Sci Ann Arbor MI 48109 USA

Modern high-dimensional datasets are often formed by acquiring samples from multiple sources having heterogeneous quality, i.e., some sources are noisier than others. Collecting data in this manner raises the following natural question: what is the best way to collect the data (i.e., how many samples should be acquired from each source) given constraints (e.g., on time or energy)? In general, the answer depends on what analysis is to be performed. In this paper, we study the foundational signal processing task of estimating underlying low-dimensional principal components. Since the resulting dataset will be high-dimensional and will have heteroscedastic noise, we focus on the recently proposed optimally weighted PCA, which is designed specifically for this setting. We develop an efficient method for designing sample acquisitions that optimize the asymptotic performance of optimally weighted PCA given resource constraints, and we illustrate the proposed method through various case studies.

关键词： Principal component analysis Noise Signal processing algorithms Costs Soft sensors software algorithms Optimization Data models Training Image sensors Heterogeneous quality large-dimensional data principal component analysis sample acquisition design

来源：评论

学校读者我要写书评

暂无评论

Interval Mean Estimation Under (ε,δ)-Local Differential Privacy

引用

IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT 2025年第2期22卷 2074-2086页

作者： Zhang, Junming Wang, Jingru Long, Shigong Li, Yanen Wang, Lun Guizhou Univ Coll Comp Sci & Technol State Key Lab Publ Big Data Guiyang 550025 Peoples R China Guizhou Polytech Construct Guiyang 550014 Peoples R China

Local differential privacy (LDP) techniques obviate the need for trust in the data collector, as they provide robust privacy guarantees against untrusted data managers while simultaneously preserving the accuracy of statistical information derived from the privatized data. As a result, these methods have garnered considerable interest and research efforts. In particular, (epsilon,delta) -LDP schemes have been utilized across a range of statistical tasks. Nonetheless, existing (epsilon,delta) -LDP mechanisms for mean estimation suffer from challenges such as elevated estimation errors and diminished data utility. To address this problem, we propose two novel (epsilon,delta) -LDP algorithms for mean estimation. Specifically, we design a one-dimensional piecewise mean estimation algorithm, which perturbs the input data into intervals, thereby reducing noise addition and enhancing both accuracy and efficiency. Building on this foundation, we extend our approach to multi-dimensional data, resulting in a multi-dimensional piecewise mean estimation algorithm. Furthermore, we conduct a theoretical analysis to derive both the variance and error bounds for the proposed algorithms. Extensive experiments conducted on real datasets demonstrate the high practicality of our algorithms for data statistical tasks, showing significant improvements in data utility.

关键词： Privacy Protection Differential privacy Accuracy Perturbation methods Noise Estimation error Probabilistic logic Big Data software algorithms Privacy protection local differential privacy mean estimation data utility

来源：评论

学校读者我要写书评

暂无评论

CURE: Simulation-Augmented Autotuning in Robotics

引用

IEEE TRANSACTIONS ON ROBOTICS 2025年 41卷 2825-2842页

作者： Hossen, Md Abir Kharade, Sonam O'Kane, Jason M. Schmerl, Bradley Garlan, David Jamshidi, Pooyan Univ South Carolina Columbia SC 29208 USA Texas A&M Univ College Stn TX 77840 USA Carnegie Mellon Univ Pittsburgh PA 15213 USA

Robotic systems are typically composed of various subsystems, such as localization and navigation, each encompassing numerous configurable components (e.g., selecting different planning algorithms). Once an algorithm has been selected for a component, its associated configuration options must be set to the appropriate values. Configuration options across the system stack interact nontrivially. Finding optimal configurations for highly configurable robots to achieve desired performance poses a significant challenge due to the interactions between configuration options across software and hardware that result in an exponentially large and complex configuration space. These challenges are further compounded by the need for transferability between different environments and robotic platforms. Data efficient optimization algorithms (e.g., Bayesian optimization) have been increasingly employed to automate the tuning of configurable parameters in cyber-physical systems. However, such optimization algorithms converge at later stages, often after exhausting the allocated budget (e.g., optimization steps, allotted time) and lacking transferability. This article proposes causal understanding and remediation for enhancing robot performance (CURE)-a method that identifies causally relevant configuration options, enabling the optimization process to operate in a reduced search space, thereby enabling faster optimization of robot performance. CURE abstracts the causal relationships between various configuration options and the robot performance objectives by learning a causal model in the source (a low-cost environment such as the Gazebo simulator) and applying the learned knowledge to perform optimization in the target (e.g., Turtlebot 3 physical robot). We demonstrate the effectiveness and transferability of CURE by conducting experiments that involve varying degrees of deployment changes in both physical robots and simulation.

关键词： Robots Optimization Transfer learning Robot sensing systems Navigation Safety Computer bugs Bayes methods Tuning software algorithms Causal inference optimization robot testing robotics and cyber-physical systems

来源：评论

学校读者我要写书评

暂无评论

Performance Benchmarking With the Austin RCS Benchmark Suite: Beyond empirical verification and validation

引用

IEEE ANTENNAS AND PROPAGATION MAGAZINE 2025年第2期67卷 95-108页

作者： Maicke, Andrew Kelley, Jon T. Yilmaz, Ali E. Univ Texas Austin Elect & Comp Engn Austin TX 78712 USA Lockheed Martin Aeronaut Ft Worth TX USA Lockheed Martin Aeronaut Palmdale CA USA

The realized performance (error-cost tradeoff) of three computational electromagnetic (CEM) methods, which use parallel algorithms on a supercomputer to predict the radar cross section (RCS) of complex targets, are quantified using the Austin RCS Benchmark Suite. The article demonstrates how modern benchmark suites can be used to evaluate CEM methods empirically and compare their performances objectively. The Austin RCS Benchmark Suite [1], [2] has recently been populated with 20 carefully selected problem sets that span a wide range in six dimensions of computational difficulty [3].

关键词： Radar cross-sections software algorithms Benchmark testing Computational electromagnetics Antennas and propagation Hardware Supercomputers software engineering Computational efficiency Parallel algorithms

来源：评论

学校读者我要写书评

暂无评论

CODEI: Resource-Efficient Task-Driven Codesign of Perception and Decision Making for Mobile Robots Applied to Autonomous Vehicles

引用

IEEE TRANSACTIONS ON ROBOTICS 2025年 41卷 2727-2748页

作者： Milojevic, Dejan Zardini, Gioele Elser, Miriam Censi, Andrea Frazzoli, Emilio Swiss Fed Inst Technol Inst Dynam Syst & Control CH-8092 Zurich Switzerland Empa Swiss Fed Labs Mat Sci & Technol Chem Energy Carriers & Vehicle Syst Lab CH-8600 Dubendorf Switzerland MIT Lab Informat & Decis Syst Cambridge MA 02139 USA

This article discusses the integration challenges and strategies for designing mobile robots, by focusing on the task-driven, optimal selection of hardware and software to balance safety, efficiency, and minimal usage of resources such as costs, energy, computational requirements, and weight. We emphasize the interplay between perception and motion planning in decision-making by introducing the concept of occupancy queries to quantify the perception requirements for sampling-based motion planners. Sensor and algorithm performance are evaluated using false negative rate and false positive rate across various factors such as geometric relationships, object properties, sensor resolution, and environmental conditions. By integrating perception requirements with perception performance, an integer linear programming approach is proposed for efficient sensor and algorithm selection and placement. This forms the basis for a codesign optimization that includes the robot body, motion planner, perception pipeline, and computing unit. We refer to this framework for solving the codesign problem of mobile robots as CODEI, short for codesign of embodied intelligence. A case study on developing an autonomous vehicle for urban scenarios provides actionable information for designers, and shows that complex tasks escalate resource demands, with task performance affecting choices of the autonomy stack. The study demonstrates that resource prioritization influences sensor choice: cameras are preferred for cost-effective and lightweight designs, while lidar sensors are chosen for better energy and computational efficiency.

关键词： Robot sensing systems Robots software Hardware Optimization software algorithms Planning Pipelines Mobile robots Decision making Codesign mobile robots sensor selection

来源：评论

学校读者我要写书评

暂无评论

Present and Future of Everyday-Use Augmented Reality Eyeglasses

引用

IEEE COMPUTER GRAPHICS AND APPLICATIONS 2025年第1期45卷 56-66页

作者： Chakravarthula, Praneeth Univ North Carolina Chapel Hill Chapel Hill NC 27599 USA

Augmented reality (AR) is emerging as the next ubiquitous wearable technology and is expected to significantly transform various industries in the near future. There has been tremendous investment in developing AR eyeglasses in recent years, including about $45 billion investment by Meta since 2021. Despite such efforts, the existing displays are very bulky in form factor and there has not yet been a socially acceptable eyeglasses-style AR display. Such wearable display eyeglasses promise to unlock enormous potential in diverse applications such as medicine, education, navigation, and many more;but until eyeglass-style AR glasses are realized, those possibilities remain only a dream. My research addresses this problem and makes progress "towards everyday-use augmented reality eyeglasses" through computational imaging, displays, and perception. My dissertation (Chakravarthula, 2021) made advances in three key and seemingly distinct areas: first, digital holography and advanced algorithms for compact, high-quality, true 3-D holographic displays;second, hardware and software for robust and comprehensive 3-D eye tracking via Purkinje Images;and third, automatic focus adjusting AR display eyeglasses for well-focused virtual and real imagery, toward potentially achieving 20/20 vision for users of all ages.

关键词： Three Dimensional Displays Navigation software algorithms Gaze Tracking Transforms Sensors Augmented Reality Investment Lenses Ubiquitous Computing Wearable Devices High Quality High Resolution Field Of View Form Factor Eye Tracking Virtual World Computer Image Tracking Accuracy Depth Of Focus Head Mounted Display Current Day Robust Tracking Wide Field Of View Object Of Attention Augmented Reality System Corneal Reflection Compactness Factor Spatial Resolution Image Quality 3 D Images Depth Perception Hologram Corrective Lenses Perception Of Quality Phase Retrieval Eye Surface Corrected Vision Refractive Error Interesting Future Direction Eye Model

来源：评论

学校读者我要写书评

暂无评论

Candidate Selection algorithms for Hybrid IP/SDN Networks With Multi-Link Failures

引用

IEEE TRANSACTIONS ON NETWORK AND SERVICE MANAGEMENT 2025年第2期22卷 1219-1231页

作者： Vuppalapati, Navya Venkatesh, T. G. Indian Inst Technol Madras Dept Elect Engn Chennai 600036 India

A longstanding problem in the context of computer networks is the issue of recovery from link failures. A promising solution to safeguard all single link failures (SLF) is through partial up gradation of certain IP routers (Internet Protocol(IP)) to software Defined Networking (SDN) switches using CSA (Candidate Selection algorithms). The resultant network comprises a combination of IP routers and SDN switches, forming a hybrid IP/SDN network. Most of the existing approaches primarily address SLFs, yet real-world scenarios may involve multiple link failures (MLF). Hence, it's crucial to investigate CSA for Hybrid IP/SDN with MLF. In this paper, we first introduce four CSA strategies tailored for hybrid IP/SDN networks with MLF. Subsequently, we present a CSA designed to accommodate network scaling or expansion. We emphasize the advantage of CSAs based on non-shortest path approach. Additionally, we propose a budget constraint-based CSA to aid in selecting suitable candidate switches within a specified budget. We evaluate the performance of hybrid IP/SDN networks employing the proposed CSAs, considering average repair path length (AMLU) and average maximum link utilization (AMLU) in the post-recovery network across diverse real-world network topologies. It will be shown that the hybrid IP/SDN with proposed CSA for MLFs will reduce the ARPL by 31.37%, 22.15%, and 18.64%, respectively, for NSF, Attmpls, and Abilene networks.

关键词： Switches IP networks Routing protocols Maintenance engineering software defined networking Tunneling Routing Network topology Topology software algorithms Candidate selection algorithm hybrid IP/SDN multi-link failures network scaling repair path length software defined networks

来源：评论

学校读者我要写书评

暂无评论

Metaheuristic algorithms in Optimizing Deep Neural Network Model for software Effort Estimation

引用

IEEE ACCESS 2021年 9卷 60309-60327页

作者： Khan, Muhammad Sufyan Jabeen, Farhana Ghouzali, Sanaa Rehman, Zobia Naz, Sheneela Abdul, Wadood COMSATS Univ Dept Comp Sci Islamabad 44000 Pakistan King Saud Univ Coll Comp & Informat Sci Dept Informat Technol Riyadh 11543 Saudi Arabia King Saud Univ Coll Comp & Informat Sci Dept Comp Engn Riyadh 11543 Saudi Arabia

Effort estimation is the most critical activity for the success of overall solution delivery in software engineering projects. In this context, the paper's main contributions to the literature on software effort estimation are twofold. First, this paper examines the application of meta-heuristic algorithms to have a logical and acceptable parametric model for software effort estimation. Secondly, to unravel the benefits of nature-inspired meta-heuristic algorithms usage in optimizing Deep Learning (DL) architectures for software effort estimation, this paper presents a Deep Neural Network (DNN) model for software effort estimation based on meta-heuristic algorithms. In this paper, Grey Wolf Optimizer (GWO) and StrawBerry (SB) meta-heuristic algorithms are applied for having a logical and acceptable parametric model for software effort estimation. To validate the performances of these two algorithms, a set of nine benchmark functions having wide dimensions is applied. Results from GWO and SB algorithms are compared with five other meta-heuristic algorithms used in literature for software effort estimation. Experimental results showed that the GWO has comprehensive superiority in terms of accuracy in estimation. The proposed DNN model (GWDNNSB) using meta-heuristic algorithms for initial weights and learning rate selection, produced better results compared to existing work on using DNN for software effort estimation.

关键词： software Estimation software algorithms Optimization Genetic algorithms Analytical models Market research software effort estimation meta-heuristic algorithms deep neural networks deep learning

来源：评论

学校读者我要写书评

暂无评论

Fast time-frequency transform algorithms and their applications to real-time software implementation of AC-3 audio codec

引用

IEEE TRANSACTIONS ON CONSUMER ELECTRONICS 1998年第2期44卷 413-423页

作者： Chen, YC Tsai, CW Wu, JL Natl Taiwan Univ Dept Comp Sci & Informat Engn Taipei 10764 Taiwan

AC-3 audio coding technology is a kind of Perceptual Audio Coder (PAC) developed by the Dolby Company. Up to 5 full-bandwidth channels and one subwoofer channel (cutoff at 120Hz) are available in AC-3 to provide multi-channel, low bit rate, and high perceptual quality of audio. This explains why AC-3 has become the audio standard of many international standards. In this paper, we focus on the real-time software implementation issues of AC-3, Two fast algorithms of the time-frequency transform, one for memory economization and the other is for time domain subsampling, are presented. Meanwhile, we will state the current performance status of our AC-3 decoder.

关键词： Time frequency analysis software algorithms Application software Psychology Bit rate Psychoacoustic models Signal processing algorithms Audio coding Decoding Multimedia systems

来源：评论

学校读者我要写书评

暂无评论

What Do AI/ML Practitioners Think About AI/ML Bias?

引用

IEEE software 2025年第1期42卷 114-118页

作者： Pant, Aastha Hoda, Rashina Turhan, Burak Tantithamthavorn, Chakkrit Monash Univ Clayton Vic 3800 Australia Monash Univ Fac Informat Technol Software Engn Clayton Vic 3800 Australia Univ Oulu Fac Informat Technol & Elect Engn Res Unit M3S Software Engn Oulu 90014 Finland

Presents a panel discussion on the topic of AI/ML Practitioners.

关键词： Artificial intelligence Ethics Companies Australia Training software engineering Collaboration software algorithms Interviews Information technology

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：