检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Vora, Jayneel Krishnan, Aditya Bouacida, Nader Shankar, Prabhu R.V. Mohapatra, Prasant Department of Computer Science University of California Davis United States Department of Electrical and Computer Engineering University of California Davis United States Department of Public Health Sciences University of California Davis Health United States Department of Computer Science University of South Florida Tampa United States

Denoising diffusion models have emerged as state-of-the-art in generative tasks across image, audio, and video domains, producing high-quality, diverse, and contextually relevant data. However, their broader adoption is limited by high computational costs and large memory footprints. Post-training quantization (PTQ) offers a promising approach to mitigate these challenges by reducing model complexity through low-bandwidth parameters. Yet, direct application of PTQ to diffusion models can degrade synthesis quality due to accumulated quantization noise across multiple denoising steps, particularly in conditional tasks like text-to-audio synthesis. This work introduces PTQ4ADM, a novel framework for quantizing audio diffusion models(ADMs). Our key contributions include (1) a coverage-driven prompt augmentation method and (2) an activation-aware calibration set generation algorithm for text-conditional ADMs. These techniques ensure comprehensive coverage of audio aspects and modalities while preserving synthesis fidelity. We validate our approach on TANGO, Make-An-Audio, and AudioLDM models for text-conditional audio generation. Extensive experiments demonstrate PTQ4ADM’s capability to reduce the model size by up to 70% while achieving synthesis quality metrics comparable to full-precision models( Copyright © 2024, The Authors. All rights reserved.

关键词： Digital elevation model

来源：评论

学校读者我要写书评

暂无评论

Identity-Focused Inference and Extraction Attacks on Diffusion Models

arXiv

引用

arXiv 2024年

The increasing reliance on diffusion models for generating synthetic images has amplified concerns about the unauthorized use of personal data, particularly facial images, in model training. In this paper, we introduce a novel identity inference framework to hold model owners accountable for including individuals’ identities in their training data. Our approach moves beyond traditional membership inference attacks by focusing on identity-level inference, providing a new perspective on data privacy violations. Through comprehensive evaluations on two facial image datasets, Labeled Faces in the Wild (LFW) and CelebA, our experiments demonstrate that the proposed membership inference attack surpasses baseline methods, achieving an attack success rate of up to 89% and an AUC-ROC of 0.91, while the identity inference attack attains 92% on LDM models trained on LFW, and the data extraction attack achieves 91.6% accuracy on DDPMs, validating the effectiveness of our approach across diffusion models. © 2024, CC BY.

关键词： Differential privacy

来源：评论

学校读者我要写书评

暂无评论

Socio-Technical Security Modelling and Simulations in Cyber-Physical Systems: Outlook on Knowledge, Perceptions, Practices, Enablers, and Barriers

引用

IET Cyber-Physical Systems: Theory and Applications 2025年第1期10卷

作者： Ani, Uchenna Daniel Al-Mhiqani, Mohammed Tuptuk, Nilufer Hailes, Stephen McKendrick Watson, Jeremy Daniel School of Computer Science and Mathematics Keele University Keele United Kingdom Department of Computer Science University of Huddersfield Huddersfield United Kingdom Department of Security and Crime Science University College London London United Kingdom Department of Computer Science University College London London United Kingdom Department of Science Technology Engineering and Public Policy University College London London United Kingdom

Socio-Technical Security Modelling and Simulation (STSec-M&S) is a technique used for reasoning and representing security viewpoints that include both the social and technical aspects of a system. It has shown great potential for improving the cybersecurity and resilience of Critical Infrastructure (CI). This study involved a multi-methods approach, consisting of a scoping literature review and a focus group workshop, conducted with stakeholder engagement from critical infrastructure stakeholders to explore their perceptions and practices regarding the use of socio-technical security modelling and simulation. The findings suggest that the current state of knowledge regarding the use and effectiveness of STSec-M&Ss approaches is limited in CI domains. Consequently, there is little application of it in existing CI systems, regardless of its recognised benefits of enabling a better understanding of CI functionalities, security goals, early and more holistic risk identifications and selection of appropriate countermeasures. The benefits of the STSec-M&S approach can be better realised by effective cross-sector communications and collaborations, team partnerships, system and approach sophistication, and better security awareness amongst others. The potential barriers that can impede such benefits include high expense for implementing the technique, low data availability and quality, regulatory compliance, and competency gaps etc. Helpful recommendations include exploring and using realistic data, validating system security models, and exploring new ways of reskilling and upskilling CI stakeholders in socio-technical security-thinking and M&S approaches to enhance cybersecurity and resilience of CIs. © 2025 The Author(s). IET Cyber-Physical Systems: Theory & Applications published by John Wiley & Sons Ltd on behalf of The Institution of engineering and Technology.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

Constrained Regularization by Denoising with Automatic Parameter Selection

arXiv

引用

arXiv 2023年

作者： Cascarano, Pasquale Benfenati, Alessandro Kamilov, Ulugbek S. Xu, Xiaojian the Department of the Arts University of Bologna Bologna40123 Italy the Department of Environmental Science and Policy University of Milan Milan20133 Italy the Department of Computer Science & Engineering Department of Electrical & System Engineering Washington University in St. Louis St. LouisMO63130 United States the Department of Electrical Engineering and Computer Science University of Michigan Ann ArborMI48109 United States

Regularization by Denoising (RED) is a well-known method for solving image restoration problems by using learned image denoisers as priors. Since the regularization parameter in the traditional RED does not have any physical interpretation, it does not provide an approach for automatic parameter selection. This letter addresses this issue by introducing the Constrained Regularization by Denoising (CRED) method that reformulates RED as a constrained optimization problem where the regularization parameter corresponds directly to the amount of noise in the measurements. The solution to the constrained problem is solved by designing an efficient method based on alternating direction method of multipliers (ADMM). Our experiments show that CRED outperforms the competing methods in terms of stability and robustness, while also achieving competitive performances in terms of image quality. © 2023, CC BY.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

Using Arduino Iot Modules As A Low Cost Environmental Research Monitoring System 1

Using Arduino Iot Modules As A Low Cost Environmental Resear...

引用

1st International Conference on computer Science and Emerging Technologies, CSET 2023

作者： Magableh, Murad Marie, Ziyad Mohamed, Rajina R. Bin Ibrahim, Mohd Hafizuddin Jusoh, Julaily Aida Kumar, Rohit Irbid National University Faculty of Business and Financial Science MIS BI and Data Analysis Department Irbid Jordan Al-Istiqlal University Faculty of Administrative Sciences and Informatics Public Administration Department Palestine College of Computing Dan Informatics Universiti Tenaga Nasional Malaysia Politeknik Kuala Terenngganu Department of Electrical Engineering Kuala Terengganu20200 Malaysia UniSZA Faculty of Informatics and Computing Malaysia School of Computer Science and Engineering Jain Deemed to Be University Bangalore India

ISBN: (纸本)9798350341737

Environmental monitoring plays a crucial role in understanding the impact of human activities on our ecosystems. However, the cost and complexity of traditional monitoring systems often hinder widespread deployment, especially in resource-constrained regions. This research paper explores the feasibility and effectiveness of employing Arduino Internet of Things (IoT) modules as a low-cost alternative for environmental research monitoring. The study focuses on designing and implementing an Arduino-based environmental monitoring system capable of gathering real-time data on key environmental parameters, such as temperature, humidity, air quality, and soil moisture. The hardware setup comprises Arduino microcontrollers interfaced with a range of environmental sensors and connectivity modules for data transmission. To assess the system's reliability and accuracy, a series of field experiments were conducted in diverse environmental settings, including urban areas, agricultural zones, and natural habitats. The collected data was compared against measurements from conventional monitoring equipment to validate the Arduino-based system's performance. The research demonstrates that the Arduino IoT modules can effectively monitor and transmit environmental data with a reasonable level of accuracy and stability, making them suitable for various environmental research applications. The low-cost nature of Arduino components and the open-source ecosystem facilitate customization and scalability, enabling researchers and organizations to develop tailored monitoring solutions for specific environmental contexts. Additionally, the paper discusses the challenges and limitations encountered during the deployment of Arduino-based monitoring systems, including power constraints, sensor calibration, and data processing techniques. Furthermore, it explores potential avenues for improvement and optimization to enhance the system's capabilities and data accuracy. In conclusion, this research

关键词： Internet of things

来源：评论

学校读者我要写书评

暂无评论

Locational Value of Distributed Energy Resources for Mitigating High Impact Events: A Case Study Based on Midwest USA

Locational Value of Distributed Energy Resources for Mitigat...

引用

IEEE General Meeting Power& Energy Society

作者： Sarangan Rajendran Adithya Melagoda Sangar Shanthanam Nirasha Bataduvaarachchi Arun-Kaarthick Manoharan Visvakumar Aravinthan Amirthagunaraj Yogarathinam Al Tamimi Charles R. Yokley Electrical and Computer Engineering Wichita State University Wichita KS USA Interdisciplinary Science Department Brookhaven National Laboratory Upton NY USA Transmission Planning and Policy Sunflower Electric Power Corporation Wichita KS USA

ISBN: (数字)9798350381832

ISBN: (纸本)9798350381849

Distributed Energy Resources (DER) aid to an electrical grid in more implicit ways beyond just supplying of power. During High Impact (HI) events caused by weather disruptions, when the grid is in a reduced state of operation, significant presence of DERs has been found to mitigate further impacts to the system. Grid operators who benefit from having DERs in their system save on system recovery costs that can be redirected back to the prosumers who contribute through DER. In this work, a methodology is introduced to value the ancillary benefits provided by the DERs during a HI event, that can be used to compensate the prosumers. The proposed method is tested on a test system with high wind energy generation, where rooftop solar PV acts as the DER, and the validity of the results are explored. The method is further expanded to include all probable variations of HI events and quantify the value of a location acting as a DER host.

关键词： Wind energy generation Measurement Costs Voltage System recovery Distributed power generation Reliability Meteorology Load modeling Load flow

来源：评论

学校读者我要写书评

暂无评论

Prediction of Voluntary Load Reduction Based on Human Behaviors During Grid Emergencies

Prediction of Voluntary Load Reduction Based on Human Behavi...

引用

Annual North-American Power Symposium

作者： Nirasha Bataduvaarachchi Demy Alexander Visvakumar Aravinthan Abigail Devereaux Al Tamimi Department of Electrical and Computer Engineering Wichita State University Wichita Kansas USA Department of Economics Wichita State University Wichita Kansas USA Transmission Planning & Policy Sunflower Electric Power Corporation Wichita Kansas USA

ISBN: (数字)9798331521035

ISBN: (纸本)9798331521042

With the growing adoption of renewable energy resources and the increasing occurrence of high-impact, low-frequency events, maintaining grid reliability has become a significant concern for many countries. Currently, there is no standardized approach for transmission-level emergency load curtailment, as grid emergencies are typically addressed from a distribution system perspective. Consumers' energy conservation efforts, also known as voluntary load reduction, can help mitigate the risk of power outages during grid emergencies. However, the existing curtailment framework does not account for voluntary reductions, leading to unfairness for consumers who proactively reduce power usage to support the grid. This oversight also results in excessive load curtailment. Therefore, predicting voluntary reductions in advance is crucial for effective grid emergency management. This paper proposes a model to predict voluntary load reduction based on collective behaviors of residential consumers. The work is carried out at the transmission level. Micro-level consumer-behaviors are modeled using agent-based modeling with numerical analysis conducted on residential households in Wichita, Kansas. The results quantify voluntary reduction based on selected household attributes and temporal environmental factors, identifying the most crucial attributes that influence significant load reductions.

关键词： Surveys Analytical models Correlation Sensitivity analysis Predictive models Emergency services Data models Numerical models Wind forecasting Load modeling

来源：评论

学校读者我要写书评

暂无评论

Realistic Sketch Face Generation via Sketch-Guided Incomplete Restoration

Realistic Sketch Face Generation via Sketch-Guided Incomplet...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Zijie Liu Chunhua Gu Ying Xie Huichen Zhang Songwen Pei School of Optical-Electrical and Computer Engineering University of Shanghai for Science and Technology Shanghai China Key Laboratory of Ministry of Public Security for Road Traffic Safety Traffic Management Research Institute of the Ministry of Public Security Wuxi China

Generating realistic sketches of human faces is an important research direction in the fields of computer vision and computer graphics. However, generating high-quality sketch faces remains challenging, especially when dealing with incomplete input sketch images. In this paper, we propose a framework based on the U-Net generative network, which progressively optimizes network performance through a three-stage training strategy. We introduce the Sketch Guidance and Completion Network (SGCN), which is capable of restoring missing features, completing blurry edges, and handling occluded regions in sketch images. By employing multiple discriminators and adversarial training strategies, we enhance the realism and preserve fine details in the generated results. Through experiments on publicly available sketch face datasets, we demonstrate significant improvements in reconstructing incomplete sketch faces using our approach. The generated images exhibit enhanced completeness, realism, and retain fine details and important facial features. Quantitative evaluations validate the effectiveness and superiority of our method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fuzzy-TOPSIS and the Urgency-Seriousness-Growth Scoring Technique to Determine Weights of Uncertainty Issues in School Reopening Factors During COVID-19

Fuzzy-TOPSIS and the Urgency-Seriousness-Growth Scoring Tech...

引用

IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)

作者： Feby Artwodini Muqtadiroh Tsuyoshi Usagawa Diana Purwitasari Riris Diana Rachmayanti Eko Mulyanto Yuniarno Supeno Mardi Susiki Nugroho Mauridhi Hery Purnomo Dept. of Electrical Engineering Dept. of Information Systems Institut Teknologi Sepuluh Nopember Surabaya Indonesia Dept. of Computer Science and Electrical Engineering Kumamoto University Kumamoto Japan Dept. of Informatics Institut Teknologi Sepuluh Nopember Surabaya Indonesia Dept. of Health Promotion and Behavioral Sciences Faculty of Public Health Universitas Airlangga Surabaya Indonesia Dept. of Electrical Engineering Dept. of Computer Engineering Institut Teknologi Sepuluh Nopember Surabaya Indonesia

The decision to open schools during a pandemic is considerable controversy because it increases the risk of a spike in COVID-19 cases more quickly. However, the decision to open schools is currently based solely on national policies, which generalize opening schools in every region, potentially leading to unintended consequences during the COVID-19 crisis. This study aims to present new decision-making to rank regions determining which schools are safe to open. By focusing on problem prioritization, we integrate the Technique for Order of Preference by Similarity to the Ideal Solution (TOPSIS) based on a fuzzy approach and the Urgency-Seriousness-Growth (USG) scoring technique, which we call FT-USG. First, we determine the weight of the expert's argument while putting special attention on the level of urgency, seriousness, and growth of COVID-19 to the school opening issue. The fuzzy number is then calculated based on the average weight generated from the USG result. Thus, the weighted normalization computations can be executed on each factor (criteria). Then, we calculate the rankings for each alternative using the Fuzzy TOPSIS method to determine which region and school can safely reopen. To provide a comprehensive ranking that serves as a basis for decision-making, we conduct a case study to show how the suggested approach is used to address the problems.

关键词：

来源：评论

学校读者我要写书评

暂无评论

High-Precision Positioning in GNSS-Challenged Environments: A LiDAR-Based Multi-Sensor Fusion Approach with 3D Digital Maps Registration

TechRxiv

引用

TechRxiv 2023年

作者： Mounier, Eslam Elhabiby, Mohamed Korenberg, Michael Noureldin, Aboelmagd The Department of Electrical and Computer Engineering Queen’s University KingstonONK7L 3N6 Canada The Department of Computer and Systems Engineering Ain Shams University Cairo11535 Egypt Micro Engineering Tech. Inc. Canada The Public Works Department Ain Shams University Cairo11535 Egypt The Department of Electrical and Computer Engineering Royal Military College of Canada KingstonONK7K 7B4 Canada

Autonomous vehicles (Avs) necessitate highly accurate and robust positioning capabilities. This calls for developing innovative algorithms that leverage the diversity of sensors and systems available in AVs to fulfill such demands across various environmental and operational conditions. This paper presents a multi-sensor positioning and navigation system that leverages cost-effective commercial-grade sensors for GNSS-challenging urban and indoor environments. The system fuses onboard motion sensor data with LiDAR point clouds registered to high-accuracy 3D digital maps to achieve sustained decimeter-level positioning. Key contributions include accurate LiDAR scans geo-referencing with motion compensation, efficient map-to-map registration, and an effective decentralized fusion. Real-world driving data from downtown Kingston, Ontario, Canada, and a high-accuracy 3D city geodatabase were used to examine the proposed methods’ performance and benefits. Results demonstrate the efficacy of the proposed technique, achieving accurate positioning with an average RMSE of 20cm horizontally and 13cm vertically, and a sustainable positioning sub-meter level of positioning accuracy 100% of the time. The proposed method was also able to sustain high precision positioning in such GNSS-denied environments with position errors of less than 50cm for 96.8% of the time and less than 30cm for 91% of the time. The performance achieved demonstrates that the proposed system is a practical option for the positioning and navigation of self-driving cars and has the potential for cooperative mapping and updating 3D city maps. © 2023, CC BY.

关键词： Autonomous vehicles

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：