检索结果-内蒙古大学图书馆

GCIM: Toward Efficient Processing of Graph Convolutional Networks in 3D-Stacked Memory

IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS 2022年第11期41卷 3579-3590页

作者： Chen, Jiaxian Lin, Yiquan Sun, Kaoyi Chen, Jiexin Ma, Chenlin Mao, Rui Wang, Yi Shenzhen Univ Coll Comp Sci & Software Engn Shenzhen 518060 Peoples R China Shenzhen Univ Shenzhen Inst Comp Sci Shenzhen 518060 Peoples R China

Graph convolutional networks (GCNs) have become a powerful deep learning approach for graph-structured data. Different from traditional neural networks such as convolutional neural networks, GCNs handle irregular input graph data, and GCNs are both computation-bound and memory-bound. How to efficiently utilize the underlying computation and memory resource becomes a critical issue. The emerging 3D-stacked computation-in-memory (CIM) architecture can reduce the data movement between computing logic and memory, thereby presenting a promising solution for the processing of GCNs. An unsolved key challenge is how to allocate GCNs to take advantage of fast near-data processing of the 3D-stacked CIM architecture. This article presents GCIM, a software-hardware co-design approach to exploit the efficient processing of GCNs on the CIM architecture. At the level of hardware design, GCIM integrates lightweight computing units near memory banks to fully exploit bank-level bandwidth and parallelism. At the level of software design, a locality-aware data mapping algorithm is proposed to partition the input graph and achieve workload balancing. GCIM is evaluated through a set of representative GCN models and standard graph datasets. The experimental results show that GCIM can significantly reduce the processing latency and data movement overhead compared with representative schemes.

关键词： software design Memory management software algorithms Systems architecture Common Information Model (computing) Hardware software 3D-stacked memory accelerator computation-in-memory (CIM) graph convolutional networks (GCNs)

来源：评论

学校读者我要写书评

暂无评论

Methods and software Tools for Converting Road Marking Line Segmentation Mask to Object Representation for Highly Automated Vehicles

Methods and Software Tools for Converting Road Marking Line ...

引用

International Russian Smart Industry Conference (SmartIndustryCon)

作者： Danil Dolgov Alexander Dolgov Imil Khairullin Faculty of Information Systems and Technologies Ulyanovsk State Technical University Ulyanovsk Russia

ISBN: (数字)9798331511241

ISBN: (纸本)9798331511258

This paper describes the development of an algorithm for converting segmentation masks derived from neural networks into ordered sets of point coordinates representing road markings for highly automated vehicles (HAVs). Several methods are considered, including the nearest neighbor method, polynomial approximation, Hough transform, and parallel contouring. Performance tests and complexity analyses are performed for each method. The parallel contour method with dynamic search for opposite points was the most effective, demonstrating linear complexity and high accuracy in approximating road marking contours. This solution ensures efficient processing of large real-time datasets, critical for HAV navigation. The study integrates the algorithm into a unified system, combining segmentation with mask-to-line conversion, enhancing autonomous driving precision.

关键词： Industries Accuracy Navigation Roads Neural networks software algorithms Transforms Approximation algorithms Polynomials Real-time systems

来源：评论

学校读者我要写书评

暂无评论

Dynamic Load Balancing on Switches of software Defined Network Managed by OpenDayLight Controller

Dynamic Load Balancing on Switches of Software Defined Netwo...

引用

International Communication Systems and Networks and Workshops, COMSNETS

作者： Manasa Kulkarni Bhargavi Goswami Muhammad Furqan Joy Paulose CHRIST (Deemed to be University) Bangalore India Queensland University of Technology Brisbane Australia

ISBN: (数字)9798331531195

ISBN: (纸本)9798331531201

In recent times, the world is becoming a global village where connectivity is a new norm irrespective of geographical location. Within corporate networks, huge setbacks are faced due to a lack of efficient resource management. Load balancing is inevitable to cater to a reliable, faster, and congestion-free communication experience for exponentially increasing online enterprises. Dynamic network resource management for high performance and low data transmission latency in a network is necessary. The major issue faced by the traditional network is that it relies on static hardware switches. software-Defined Network approaches paved a way to overcome the limitations of traditional networks. This research proposes the Dynamic Load Balancing Algorithm for software-Defined Networks to utilize network resources optimally. The major function of the proposed algorithm is to determine alternative paths and further distribute the incoming and outgoing network flows to achieve optimum network resource utilization with faster traffic flow completion. The experiment is performed with the OpenDayLight Controller on the Mininet simulator, which emulates the network with the novel scheme. The results prove that the proposed solution has accomplished the benchmarks of optimum throughput, reduced redundancy, and reduced flow completion time.

关键词： Heuristic algorithms software algorithms Redundancy Packet loss Load management Throughput Dynamic scheduling Control systems Reliability software defined networking

来源：评论

学校读者我要写书评

暂无评论

Bridging AI, Robotics, and software Engineering: An Interdisciplinary Approach for Learning Emerging Technologies

Bridging AI, Robotics, and Software Engineering: An Interdis...

引用

IEEE Education Engineering (EDUCON)

作者： Lorena B. Martínez Elizalde Carlos Astengo Noguez Maria Raquel Landa Cavazos Luis Ricardo Salgado Garza Department of Computer Science Tec de Monterrey Monterrey Mexico Department of Sciences Tec de Monterrey Monterrey Mexico

ISBN: (数字)9798331539498

ISBN: (纸本)9798331539504

As the rapid evolution of emerging technologies continues to reshape industries, the need for an interdisciplinary approach to learning has become critical. Artificial intelligence (AI), robotics, and software engineering are key fields driving innovation, yet they are often taught in isolation. This paper advocates for a comprehensive, interdisciplinary learning model that integrates AI, robotics, and software engineering to prepare students and professionals for the demands of the future workforce. By blending theoretical knowledge with handson experience across these domains, the proposed approach fosters a deeper understanding of how these technologies intersect and support each other in real-world applications. The main problem addressed in this article is the fragmentation of higher education curricula, where AI, robotics, and software engineering are traditionally taught as separate subjects. This separation limits students' ability to grasp how these fields can collaborate to solve complex problems, leading to a knowledge gap that hinders their ability to innovate in an interconnected technological landscape. The proposed interdisciplinary model aims to close this gap by providing a holistic education that integrates key principles from all three domains. This approach aims to equip students with a versatile skill set that is adaptable to the rapidly changing technological environment. Through a combination of lectures, lab work, and project-based learning, students will explore the synergies between AI algorithms, robotic systems, and software development processes. The curriculum emphasizes the importance of collaboration between these fields, demonstrating how AI can enhance robotic autonomy, how software engineering principles ensure the scalability of AI models, and how robotics offers practical applications for these technologies. The reasoning behind this interdisciplinary model is grounded in the idea that emerging technologies do not exist in silos.

关键词： Industries Adaptation models Technological innovation Service robots software algorithms Teamwork Artificial intelligence Engineering education software engineering software development management

来源：评论

学校读者我要写书评

暂无评论

software tool-set for automated quantum system identification and device bring up

Software tool-set for automated quantum system identificatio...

引用

International Communication Systems and Networks and Workshops, COMSNETS

作者： Anurag Saha Roy Kevin Pack Nicolas Wittler Shai Machnes Peter Grünberg Institute – Quantum Computing Analytics (PGI 12) Forschungszentrum Jülich Jülich Germany Theoretical Physics Saarland University Saarbrücken Germany Qruise GmbH Saarbrücken Germany

ISBN: (数字)9798331531195

ISBN: (纸本)9798331531201

We present a software tool-set which combines the theoretical, optimal control view of quantum devices with the practical operation and characterization tasks required for quantum computing. In the same framework, we perform model-based simulations to create control schemes, calibrate these controls in a closed-loop with the device (or in this demo —by emulating the experimental process) and finally improve the system model through minimization of the mismatch between simulation and experiment, resulting in a digital twin of the device. The model based simulator is implemented using TensorFlow, for numeric efficiency, scalability and to make use of automatic differentiation, which enables gradient-based optimization for arbitrary models and control schemes. Optimizations are carried out with a collection of state-of-the-art algorithms originated in the field of machine learning. All of this comes with a user-friendly Qiskit interface, which allows end-users to easily simulate their quantum circuits on a high-fidelity differentiable physics simulator.

关键词： Quantum system Computational modeling Scalability software algorithms Process control software Numerical models Object recognition Quantum circuit Optimization

来源：评论

学校读者我要写书评

暂无评论

A Soft Constrained MPC Formulation Enabling Learning From Trajectories With Constraint Violations

引用

IEEE CONTROL SYSTEMS LETTERS 2022年 6卷 980-985页

作者： Wabersich, Kim P. Krishnadas, Raamadaas Zeilinger, Melanie N. Swiss Fed Inst Technol Inst Dynam Syst & Control CH-8053 Zurich Switzerland

In practical model predictive control (MPC) implementations, constraints on the states are typically softened to ensure feasibility despite unmodeled disturbances. In this work, we propose a soft constrained MPC formulation supporting polytopic terminal sets in half-space and vertex representation, which significantly increases the feasible set while maintaining asymptotic stability in case of constraint violations. The proposed formulation allows for leveraging system trajectories that violate state constraints to iteratively improve the MPC controller's performance. To this end, we apply convex optimization techniques to obtain a data-driven terminal cost and set, which result in a quadratic MPC problem.

关键词： Cost function Asymptotic stability Trajectory Stability analysis Softening Predictive control software algorithms Predictive control for linear systems constrained control iterative learning control

来源：评论

学校读者我要写书评

暂无评论

Method and Process Mining Based software for ECG Analysis and Arrhythmia Detection

Method and Process Mining Based Software for ECG Analysis an...

引用

Digital Signal Processing and its Applications (DSPA), International Conference on

作者： Kseniya Demushkina Andrey Kuzmin Information and Computing Systems Penza state University Penza Russia

ISBN: (数字)9798331518752

ISBN: (纸本)9798331518769

Currently, the use of neural network based solutions in medical decision support systems remains popular. In order to overcome the hidden decision-making process called “black box”, the authors propose to use the “white box” principle based on the process mining approach. The aim of the work is to develop and implement a process mining based method of electrocardiosignal analysis and arrhythmia detection. Pan-Tompkins algorithm for electrocardiosignal signal analysis and ProM 6.11 framework for software implementation were used. As a results a new intelligent process mining based electrocardiosignal analysis method is proposed. It consists of conversion of digital data of cardiac signal into event log, construction of Petri net and construction of diagnostic decision tree. The original plugin generates the event log in the .xes format using Pan-Tompkins algorithm. ProM 6.11 AlphaMiner plugin generates a Petri net that reflects the relationship of events in the log. The resulting event log allows building a diagnostic decision tree using ProM 6.11 discovery of the process data-flow plugin. The study demonstrates the applicability of the proposed method using normal and arrhythmic cardiac signal records from PhysioNet database. The proposed solution on the base of process mining technology and ProM 6.11 framework allows implementing software for transparent analysis of cardiac signals data for medical decision support systems. The next step is to conduct statistical assessments of the effectiveness of the proposed solution and to expand the number of informative parameters used for analysis.

关键词： Process mining Decision support systems Arrhythmia PROM software algorithms Signal processing algorithms Electrocardiography software Decision trees Medical diagnostic imaging

来源：评论

学校读者我要写书评

暂无评论

AIPD: A Novel API Interception-Based Patch Deployment Method to Boost software Security

AIPD: A Novel API Interception-Based Patch Deployment Method...

引用

2025 2nd International Conference on algorithms, software Engineering and Network Security (ASENS)

作者： Hongyao Li Xianglong Kong Mingqi Qiao Xin Chen Zhizhong Zhou Xiao Huang School of Cyber Science and Engineering Southeast University Nanjing China Purple Mountain Laboratories Nanjing China

ISBN: (数字)9798331520298

ISBN: (纸本)9798331520304

Open source software (OSS) vulnerabilities are crucial for ensuring the security of software systems. Existing methods typically require application developers to patch their own operational environment clients, overlooking the security responsibilities that should be assumed by the operational environment providers. This often leads to failed patch deployments and affects the stability of the software systems. To address this issue, this paper proposes AIPD, a novel method for completing patch deployment at the manufacturing side. AIPD does not require modification of the OSS source code; instead, it extracts the core APIs involved in the vulnerability fixes and their dependencies from the source version, dynamically intercepts them in the target version's system environment, and replaces them with the patched API for secure updates. Experiments show that AIPD achieves a 91.35% patching accuracy on a dataset of 104 vulnerability patches containing 10 Python packages. In compatibility testing for the patch deployment environment, the pass rate for the relevant APIs reached 97.69%, effectively reducing the compatibility risks of the system environment.

关键词： Source coding software algorithms Network security Maintenance engineering software systems Manufacturing Open source software Python Testing software engineering

来源：评论

学校读者我要写书评

暂无评论

Boosting software Quality: A Machine Learning Technique for Predicting Defects

Boosting Software Quality: A Machine Learning Technique for ...

引用

Mobile Computing and Sustainable Informatics (ICMCSI), International Conference on

作者： Akilan E. Daniel Alwin Infant P Jainish G R Computer science & Engineering Panimalar Engineering College Chennai City Campus Chennai India Department of Computer Science and Engineering Panimalar Engineering College Chennai India Computer Science and Engineering Loyola ICAM College of Engineering & Technology Chennai India

ISBN: (数字)9798331522667

ISBN: (纸本)9798331522674

In order to increase automation, strengthen decision- making, and improve efficiency while reducing costs and maintaining precision, the development of intelligent machinery is becoming more and more dependent on the integration of data, embedded software, and modern technologies. Reliability of embedded software becomes essential for guaranteeing high system performance and an improved user experience as technology and hardware merge to build smart devices. This work explores methods to quantify software reliability, which measures the likelihood of residual failures in a system. The objective is to build a statistical relationship between quality parameters and product or process metrics through the analysis of software metrics produced from development data. These indicators aid in evaluating the software's dependability and performance, enabling developers to spot and fix possible problems early on. Using pattern recognition algorithms to pinpoint software's fault-prone areas is a crucial strategy for enhancing software reliability prediction. These algorithms enable developers to focus attention on software areas that have a higher likelihood of failure by identifying patterns in the product's code or structure that suggest increased risk. In summary, these methods focus on enhancing the reliability of intelligent machines by equipping developers with advanced tools to reduce defects, improve dependability, and ensure that smart systems perform optimally. This approach aligns with modern trends in using machine learning and data-driven models to tackle complex technological challenges.

关键词： Computational modeling software algorithms software quality Predictive models Prediction algorithms Data models software reliability Decision trees software measurement Embedded software

来源：评论

学校读者我要写书评

暂无评论

AI in software Patenting: Jurisprudential Analysis under the European Patent Convention, UK, and US Legal Frameworks

AI in Software Patenting: Jurisprudential Analysis under the...

引用

Advances in Computing, Communication and Networking (ICAC2N), International Conference on

作者： Ahtshamuddin Ansari Rajneesh Arshi Ansari Mithilesh Kumar Yadav Vishva Gaurav Ashutesh Anand College of Law IIMT University Meerut U.P. School of Arts and Humanities IIMT University Meerut U.P.

ISBN: (数字)9798350356816

ISBN: (纸本)9798350356823

Artificial intelligence in software patenting presents unique problems and creates opportunities across a wide variety of legal jurisdictions. Drawing on jurisprudential analysis under the European Patent Convention, the United Kingdom, and the legal frameworks of the United States of America, this paper sheds lighter on these intricacies. It provides, among other things, the requirements of patentability with an emphasis on the inventive step and novelty in AI-driven inventions. Most importantly, it signals divergences of the legal interpretation and practice of such within these regions and gives space to such polemics as questions about algorithm patentability or human contribution and all this means for innovation policy. The comparative legal analysis herein offers insight into a fast-changing landscape of AI software patenting. Finally, it provides recommendations on ways and means to harmonize further laws regulating patents in light of rapid technological development in AI.

关键词： Training Patents Technological innovation Law software algorithms Europe software Artificial intelligence Protection Standards

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：