检索结果-内蒙古大学图书馆

arXiv 2022年

作者： Dantas, Joao P.A. Costa, Andre N. Medeiros, Felipe L.L. Geraldo, Diego Maximo, Marcos R.O.A. Yoneyama, Takashi Decision Support Systems Subdivision Institute for Advanced Studies Trevo Cel Av Jose A. A. do Amarante 1 Putim SP Sao Jose dos Campos12228-001 Brazil Computer Science Division Aeronautics Institute of Technology Praca Marechal Eduardo Gomes 50 Vila das Acacias SP Sao Jose dos Campos12228-900 Brazil Electronic Engineering Division Aeronautics Institute of Technology Praca Marechal Eduardo Gomes 50 Vila das Acacias SP Sao Jose dos Campos12228-900 Brazil

This work compares supervised machine learning methods using reliable data from constructive simulations to estimate the most effective moment for launching missiles during air combat. We employed resampling techniques to improve the predictive model, analyzing accuracy, precision, recall, and f1-score. Indeed, we could identify the remarkable performance of the models based on decision trees and the significant sensitivity of other algorithms to resampling techniques. The models with the best f1-score brought values of 0.379 and 0.465 without and with the resampling technique, respectively, which is an increase of 22.69%. Thus, if desirable, resampling techniques can improve the model's recall and f1-score with a slight decline in accuracy and precision. Therefore, through data obtained through constructive simulations, it is possible to develop decision support tools based on machine learning models, which may improve the flight quality in BVR air combat, increasing the effectiveness of offensive missions to hit a particular target. Copyright © 2022, The Authors. All rights reserved.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

A REINFORCEMENT LEARNING METHOD FOR ENVIRONMENTS WITH STOCHASTIC VARIABLES: POST-DECISION PROXIMAL POLICY OPTIMIZATION WITH DUAL CRITIC NETWORKS

arXiv

引用

arXiv 2025年

作者： Felizardo, Leonardo Kanashiro Fadda, Edoardo Brandimarte, Paolo Del-Moral-Hernandez, Emilio Vasconcelos Nascimento, Mariá Cristina Division of Computer Science Instituto Tecnológico de Aeronáutica Brazil Department of Mathematical Sciences Politecnico di Torino Italy Escola Politécnica Universidade de São Paulo Brazil Division of Computer Science Instituto Tecnológico de Aeronáutica São José dos Campos Brazil

This paper presents Post-Decision Proximal Policy Optimization (PDPPO), a novel variation of the leading deep reinforcement learning method, Proximal Policy Optimization (PPO). The PDPPO state transition process is divided into two steps: a deterministic step resulting in the post-decision state and a stochastic step leading to the next state. Our approach incorporates post-decision states and dual critics to reduce the problem’s dimensionality and enhance the accuracy of value function estimation. Lot-sizing is a mixed integer programming problem for which we exemplify such dynamics. The objective of lot-sizing is to optimize production, delivery fulfillment, and inventory levels in uncertain demand and cost parameters. This paper evaluates the performance of PDPPO across various environments and configurations. Notably, PDPPO with a dual critic architecture achieves nearly double the maximum reward of vanilla PPO in specific scenarios, requiring fewer episode iterations and demonstrating faster and more consistent learning across different initializations. On average, PDPPO outperforms PPO in environments with a stochastic component in the state transition. These results support the benefits of using a post-decision state. Integrating this post-decision state in the value function approximation leads to more informed and efficient learning in high-dimensional and stochastic environments. Copyright © 2025, The Authors. All rights reserved.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

SUN-stealth UAV networks: A cooperative local model to mitigate UAVs exposure to threats

SUN-stealth UAV networks: A cooperative local model to mitig...

引用

2019 IEEE International Conference on Systems, Man and Cybernetics, SMC 2019

作者： Borges, Nicolas Pereira Ghedini, Cinara Guellner Ribeiro, Carlos Henrique Costa Computer Science Division at Aeronautics Institute of Technology São José dos Campos SP Brazil

ISBN: (纸本)9781728145693

Cooperative unmanned aerial vehicles (UAVs) are suitable for complex missions in denied environments. In scenarios where denial is potentially hostile, the use of stealth approaches is essential, as it allows such missions to be conducted in the proximity of threatening radars and weapons. We propose a model that allows autonomous UAVs to dynamically and cooperatively reduce their exposure to threats by adopting a stealth policy, in which they turn off their communication radios once a threat is detected. We named this approach of cooperative model as Stealth UAV Networks (SUN). Our model also encompasses control laws for connectivity maintenance, collision avoidance between UAVs, flight path planning, coverage area and robustness to communication failures improvement. We validated the approach with different combinations of control laws, number of UAVs and scenario hostility. The exposure of UAVs to threats decreased by at least 61% and at most 96% when the stealth policy is active, considering all simulation settings. Thus, highlighting that the stealth policy can be successfully used in missions that demand the increase of coverage area rate or robustness to failures. © 2019 IEEE.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Future Perspectives in Healthcare: An Analysis of Augmented Reality and Spatial Computing in Hospital Environments

引用

Procedia computer science 2024年 238卷 932-937页

作者： Carlos Alves José Luís Reis José Paulo Marques dos Santos Luís Mendes Gomes José Machado ALGORITMI/LASI Research Centre Escola de Engenharia - Universidade do Minho Campus Azurém 4800-058 Guimarães Portugal Department of Business Administration - University of Maia Av. Carlos de Oliveira Campos 4475-690 Maia Portugal Artificial Intelligence and Computer Science Laboratory (LIACC) - University of Porto Rua Dr. Roberto Frias 4200-465 Porto Portugal LASI - Intelligent Systems Associate Laboratory - Campus Azurém 4800-058 Guimarães Portugal ALGORITMI/LASI Universidade dos Açores 9500-321 Ponta Delgada Portugal

This article investigates the applicability of Augmented Reality (AR) and Spatial Computing in hospital environments, exploring how these technologies can enhance interaction between patients and healthcare professionals as well as the efficiency of medical procedures. Through a systematic literature review using the PRISMA methodology, relevant articles highlighting the potential of these technologies in the hospital setting were selected and analysed. The results suggest a growing adoption of AR and Spatial Computing, driven by the advancement of 5G technology, with the potential to improve diagnostic accuracy, the effectiveness of surgical procedures, and the educational experience of healthcare professionals. The study concludes that the integration of these emerging technologies promises to transform hospital environments, emphasising the need to overcome technical, ethical, and privacy challenges to maximise their positive impact on health.

关键词： Augmented Reality AR Spatial Computing Helathcare

来源：评论

学校读者我要写书评

暂无评论

Push Recovery Strategies through Deep Reinforcement Learning

Push Recovery Strategies through Deep Reinforcement Learning

引用

IEEE Latin American Robotics Symposium, LARS

作者： Dicksiano Carvalho Melo Marcos R. O. A. Máximo Adilson Marques da Cunha Autonomous Computational Systems Lab (LAB-SCA) Computer Science Division Aeronautics Institute of Technology São José dos Campos SP Brazil Computer Science Division Aeronautics Institute of Technology SãoJosé dos Campos SP Brazil

ISBN: (数字)9780738111537

ISBN: (纸本)9781665419857

This work implements a Push Recovery controller that improves a walking engine used by a humanoid simulated agent from RoboCup 3D Soccer Simulation League. Therefore, we achieve a movement policy that allows the agents to re-establish their own balance when disturbed. Using the Proximal Policy Optimization algorithm, this work achieved an expert policy, represented by a Deep Neural Network, which is compatible with a Zero Moment Point walking engine. In contrast with other works, the reward signal used is much more agnostic, in a way that the learning agent was not aware of the physical constraints related to the problem. The agent was just aware that it should avoid failing in order to accumulate more reward. The proposed method is also sample efficient, as it is able to learn a human inspired behavior in just a few hours.

关键词： Robots Hip Humanoid robots Legged locomotion Robot sensing systems Engines Torque

来源：评论

学校读者我要写书评

暂无评论

Transition from bounded to unbounded energy in a time-dependent billiard

引用

Physical Review E 2025年第5期111卷 054215-054215页

作者： Anne Kétri P. da Fonseca Felipe Augusto O. Silveira Célia M. Kuwana Diego F. M. Oliveira Edson D. Leonel Departamento de Física UNESP–Universidade Estadual Paulista Avenida 24A 1515 Bela Vista Rio Claro 13506-900 São Paulo Brazil Universidade Federal de São Paulo Instituto de Ciência e Tecnologia Avenida Cesare Mansueto Giulio Lattes 1201 Eugênio de Mello São José dos Campos 12247-014 São Paulo Brazil School of Electrical Engineering and Computer Science University of North Dakota Avenue Stop 8357Grand Forks North Dakota 58202 USA

We revisit a time-dependent, oval-shaped billiard to investigate a phase transition from bounded to unbounded energy growth. The introduction of inelastic collisions between the particle and the boundary limit the unbounded energy increase after the crossover iteration observed for the conservative dynamics. The central phenomenology uses a probability distribution that satisfies the system's boundary conditions to solve the diffusion equation. From this, we express the observables and, using a set of scaling hypotheses and a generalized homogeneous function, obtain a relation between the critical exponents, leading to a scaling law validated through numerical simulations and analytical methods. This transition displays properties similar to continuous phase transitions in statistical mechanics, including scale invariance, interrelated critical exponents governed by scaling laws, and an order parameter (susceptibility) approaching zero (infinity) at the transition. Furthermore, the system exhibits a variation in velocity between collisions mediated by control parameters, known as elementary excitations, which facilitate particle diffusion. The system also does not present topological defects that could alter the probability distribution function.

关键词： Iterative methods

来源：评论

学校读者我要写书评

暂无评论

iRED: A disaggregated P4-AQM fully implemented in programmable data plane hardware

arXiv

引用

arXiv 2023年

作者： de Almeida, Leandro C. Pasquini, Rafael Papagianni, Chrysa Verdi, Fábio L. Academic Unit of Informatics Federal Institute of Paraíba Av. Primeiro de Maio PB João Pessoa58015-435 Brazil Faculty of Computing Federal University of Uberlândia Av. João Naves de Ávila MG Uberlândia38400-902 Brazil Faculty of Science University of Amsterdam Amsterdam1098 XH Netherlands Department of Computer Science Federal University of São Carlos Sorocaba Campus Rodovia João Leme dos Santos SP Sorocaba18052-780 Brazil

Routers employ queues to temporarily hold packets when the scheduler cannot immediately process them. Congestion occurs when the arrival rate of packets exceeds the processing capacity, leading to increased queueing delay. Over time, Active Queue Management (AQM) strategies have focused on directly draining packets from queues to alleviate congestion and reduce queuing delay. On Programmable Data Plane (PDP) hardware, AQMs traditionally reside in the Egress pipeline due to the availability of queue delay information there. We argue that this approach wastes the router’s resources because the dropped packet has already consumed the entire pipeline of the device. In this work, we propose ingress Random Early Detection (iRED), a more efficient approach that addresses the Egress drop problem. iRED is a disaggregated P4-AQM fully implemented in programmable data plane hardware and also supports Low Latency, Low Loss, and Scalable Throughput (L4S) framework, saving device pipeline resources by dropping packets in the Ingress block. To evaluate iRED, we conducted three experiments using a Tofino2 programmable switch: i) An in-depth analysis of state-of-the-art AQMs on PDP hardware, using 12 different network configurations varying in bandwidth, Round-Trip Time (RTT), and Maximum Transmission Unit (MTU). The results demonstrate that iRED can significantly reduce router resource consumption, with up to a 10x reduction in memory usage, 12x fewer processing cycles, and 8x less power consumption for the same traffic load;ii) A performance evaluation regarding the L4S framework. The results prove that iRED achieves fairness in bandwidth usage for different types of traffic (classic and scalable);iii) A comprehensive analysis of the Quality of Service (QoS) in a real setup of a Dynamic Adaptive Streaming over HTTP (DASH) technology. iRED demonstrated up to a 2.34x improvement in Frames Per Second (FPS) and a 4.77x increase in the video player buffer fill. © 2023, CC BY.

关键词： Drops

来源：评论

学校读者我要写书评

暂无评论

Data analysis platform for the optimization of employability in technological profiles 17th

Data analysis platform for the optimization of employability...

引用

17th International Conference on Practical Applications of Agents and Multi-Agent Systems, PAAMS 2019

作者： Chamoso, Pablo González-Briones, Alfonso García-Peñalvo, Francisco José Department of Computer Science and Automation University of Salamanca Plaza de los Caídos s/n Salamanca Spain

ISBN: (纸本)9783030242985

The information technology (IT) sector has not only created the largest number of jobs but has also contributed to greater employability in other sectors. Thanks to major advances in computing, the analysis of large volumes of data for extraction of information unknown a priori, has become a trend in all sectors and its benefits and advantages are unquestionable. The aim of this research work is to extract knowledge from employability information for market trend analysis and use this knowledge to adapt the user’s search for work to their profile and guide them accordingly. To this end, existing employability and training information will be retrieved, analysed and a platform will be created to allow the user to easily visualise the results, enabling users with no knowledge of data analysis to perform studies based on machine learning. © Springer Nature Switzerland AG 2019.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

Tenant-Oriented Resource optimization for Cloud Network Slicing with Performance Guarantees

Tenant-Oriented Resource optimization for Cloud Network Slic...

引用

IEEE Conference on Network Softwarization (NetSoft)

作者： Lucian Beraldo Angelos Pentelas Fábio Luciano Verdi Panagiotis Papadimitriou Cesar A. C. Marcondes Federal University of São Carlos (UFSCar) São Carlos SP Brazil University of Macedonia Thessaloniki Greece Computer Science Division Aeronautics Institute of Technology (ITA) São José dos Campos SP Brazil

Cloud Network Slicing (CNS), emerging alongside the 5G mobile network, comprises a paradigm shift in the way networks are provisioned, managed, and operated. Fundamentally, CNS fosters the deployment of a multitude of modern applications, e.g., virtual and augmented reality, 4K video streaming, and autonomous vehicles, which require ultra-low latency, high bandwidth consumption, or both. Slicing promotes the realization of such services through the allocation of computing and network resource bundles, which, as CNS mandates, are isolated from the rest of the network. Typically, such resources are arranged into wide geographical areas (e.g., into multiple countries or even continents), which implies that it is possible to allocate from multiple infrastructure providers. This exacerbates the already challenging problem of maximizing resource allocation efficiency, a feature commonly addressed by CNS *** this respect, we study the optimized embedding of slices across multiple domains. Therefore, we account for slices as a collection of computing and network parts. Given specific resource requirements from slice tenants and potentially multiple offers per slice part, we model the problem as a Mixed Integer Linear Program (MILP). We further design two heuristic algorithms in order to mitigate the complex intricacies that would be perceptible in large problem instances. Our evaluation results, based on a simulation environment aligned with the NECOS slicing architecture, indicate that the MILP approach yields better efficiency compared to both heuristics, with respect to client expenditure with a fair amount of performance parameters in an adequate execution time. Our main contribution lies in the optimization methods based on the split and combine approach, integrated into the NECOS CNS architecture.

关键词： Cloud computing Waste materials 5G mobile communication Network slicing Optimization methods computer architecture Workstations

来源：评论

学校读者我要写书评

暂无评论

Watermarking of computer Generated Imagery : A Review 3

Watermarking of Computer Generated Imagery : A Review

引用

3rd IEEE International Conference on Electrical, computer and Communication Technologies: computer science and Engineering, ICECCT 2019

作者： Kumar, H.B. Basanth Chennamma, H.R. DoS in Computer Science Pooja Bhagavat Memorial Mahajana Education Centre Mysore India Department of MCA JSS Science and Technology University Mysore India

ISBN: (纸本)9781538681572

computer graphics is enormously used in newspapers, magazines, advertisements, video games, architecture design and modeling, satellite images, maps and many more fields. Recent developments in information technology enable the user to produce, modify and distribute the graphics content more easily. Nowadays, protection of graphics content from being illegal usage by the unauthorized users has become a major concern. Various watermarking techniques have been proposed by the researchers for copyright protection and authentication of graphics content by embedding the secret information into it which is imperceptible to the users. This paper presents an overview of existing watermarking schemes used for copyright protection and to verify the authenticity and integrity of the graphic contents. © 2019 IEEE.

关键词： Animation

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：