检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Shan, Jianhua Sun, Yuhao Zhang, Shixin Sun, Fuchun Chen, Zixi Shen, Zirong Stefanini, Cesare Yang, Yiyong Luo, Shan Fang, Bin School of Mechanical Engineering Anhui University of Technology China Institute for Artificial Intelligence Department of Computer Science and Technology Beijing National Research Center for Information Science and Technology Tsinghua University Beijing100084 China Beijing100083 China Biorobotics Institute The Department of Excellence in Robotics and AI Scuola Superiore Sant’Anna Pisa56127 Italy Department of Engineering King’s College London LondonWC2R 2LS United Kingdom Zhili College Tsinghua University Beijing100084 China

Deformable object manipulation is a challenging problem due to its complex deformable properties. With the development of artificial intelligence, learning-based methods have shown outstanding performance in the manipulation of robots. Previous works have investigated the manipulation of deformable objects via Reinforcement Learning (RL) in simulation. However, they approximate object deformation with particles, and particle states are employed as observations, which are not available in reality. To address these issues, we propose a novel approach utilizing vision-Based Tactile Sensors (VBTSs) as the end-effector in simulation to produce observations like relative position, squeezed area, and object contour, which are transferable to real robots. However, this makes contact simulation more complex due to the gel layer of VBTS is also a deformable object. Existing simulation methods of the vision-based tactile sensor can only simulate elastic deformation, while the simulation of plastic and elastoplastic deformation is poor. This must be overcome for a more realistic contact simulation between deformable objects. In this work, we build a new contact simulation environment for deformable objects including elastic, plastic, and elastoplastic deformation. We utilize RL strategies to train agents in the simulation, and expert demonstrations are applied for challenging tasks. To achieve simulation-to-real-world (sim-to-real) transfer, transferable observations like relative position, squeezed area, and object contour are applied in RL training. Also, we build a real experimental platform, including a VBTS, to complete the sim-to-real work and robustness testing. Leveraging the developed simulation and real experiment setup, a benchmark has been created for contact simulation and manipulation learning of deformable objects with VBTS. Our work innovatively proposes a strategy that employs high-resolution VBTSs in contact simulation and manipulation of deformable objects. W

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Open-Pose 3d Zero-Shot Learning: Benchmark and Challenges

SSRN

引用

SSRN 2024年

作者： Zhao, Weiguang Yang, Guanyu Zhang, Rui Jiang, Chenru Yang, Chaolong Yan, Yuyao Hussain, Amir Huang, Kaizhu Department of Computer Science University of Liverpool LiverpoolL69 7ZX United Kingdom Department of Foundational Mathematics Xi’an Jiaotong-Liverpool University Suzhou215123 China Department of Mechatronics and Robotics Xi’an Jiaotong-Liverpool University Suzhou215123 China School of Robotic Xi’an Jiaotong-Liverpool University Suzhou215123 China School of Computing Edinburgh Napier University EdinburghEH11 4BN United Kingdom Data Science Research Center Duke Kunshan University Kunshan215316 China

With the explosive 3D data growth, the urgency of utilizing zero-shot learning to facilitate data labeling becomes evident. Recently, methods transferring language or language-image pre-training models like Contrastive Language-Image Pre-training (CLIP) to 3D vision have made significant progress in the 3D zero-shot classification task. These methods primarily focus on 3D object classification with an aligned pose;such a setting is, however, rather restrictive, which overlooks the recognition of 3D objects with open poses typically encountered in real-world scenarios, such as an overturned chair or a lying teddy bear. To this end, we propose a more realistic and challenging scenario named open-pose 3D zero-shot classification, focusing on the recognition of 3D objects regardless of their orientation. First, we revisit the current research on 3D zero-shot classification, and propose two benchmark datasets specifically designed for the open-pose setting. We empirically validate many of the most popular methods in the proposed open-pose benchmark. Our investigations reveal that most current 3D zero-shot classification models suffer from poor performance, indicating a substantial exploration room towards the new direction. Furthermore, we study a concise pipeline with an iterative angle refinement mechanism that automatically optimizes one ideal angle to classify these open-pose 3D objects. In particular, to make validation more compelling and not just limited to existing CLIP-based methods, we also pioneer the exploration of knowledge transfer based on Diffusion models. While the proposed solutions can serve as a new benchmark for open-pose 3D zero-shot classification, we discuss the complexities and challenges of this scenario that remain for further research development. The code is available publicly at https://***/weiguangzhao/Diff-OP3D. © 2024, The Authors. All rights reserved.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Dark vision Assistive Application for Deaf-Blind People

Dark Vision Assistive Application for Deaf-Blind People

引用

Computing and Information Technology (ICCIT-1441), 2020 International Conference on

作者： Prithi Samuel Rajesh Kumar Dhanaraj Balamurugan Balusamy Narmatha C Department of Computational Intelligence SRM Institute of Science and Technology Chennai India Symbiosis Institute of Computer Studies and Research (SICSR) Symbiosis International (Deemed University) Pune India Associate Dean- Students Shiv Nadar University Noida India Industrial Innovation & Robotics Center Faculty of Computers and Information Technology University of Tabuk Tabuk City Kingdom of Saudi Arabia

For visually impaired, deaf-blind people, both communication and navigation are difficult activities and indoor navigation is becoming increasingly harder for them. It is even more difficult for deaf-blind people than it is for non-visually impaired people. People with visual impairments or deaf- blindness frequently rely on external support systems to make judgments, such as trained canines, humans, or specialized equipment. As a result, deaf-blind people are in desperate need of an assistive application that allows blind people to navigate and communicate freely without relying on others. The goal of the application is to create a more adaptable and cost-effective system for deaf-blind and visually impaired people. Through both outdoor and interior applications, as well as walking in even unfamiliar locations without relying on others.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Using system dynamics to inform scenario planning: Application to the Souss-Massa basin, Morocco

引用

Journal of Urban Management 2025年

作者： Guemouria, Ayoub Chehbouni, Abdelghani Belaqziz, Salwa Dhiba, Driss Bouchaou, Lhoussaine International Water Research Institute Mohammed VI Polytechnic University Benguerir 43150 Morocco Center for Remote Sensing Applications Mohammed VI Polytechnic University Benguerir 43150 Morocco Institut de Recherche pour le Développement Unité Mixte de Recherche Centre d’études Spatiales de la Biosphère Toulouse 31400 France Laboratory of Computer Systems and Vision Department of Computer Science Faculty of Science Ibn Zohr University Agadir 80000 Morocco Laboratory of Applied Geology and Geo-Environment Faculty of Science Ibn Zohr University Agadir 80000 Morocco

The watershed represents a holistic system whose poor understanding of its multiple subsystems can lead to a pronounced water scarcity. This study aims to develop an innovative technique for managing water resources within the Souss-Massa watershed. It uses the System Dynamics (SD) methodology to analyze the interplay among the factors involved in water supply and demand. The results show that under the Business As Usual (BAU) scenario, water sustainability in this watershed is not assured. Groundwater drawdown (GWD) will increase significantly, with an estimated average decrease of −337 Mm³ for the period 2022 to 2050. To remedy this critical situation, several simulations were developed, each representing a distinct scenario. Scenario 1 improves irrigation efficiency by 10%, while scenario 2 achieves a 20% improvement. Scenario 3 builds on scenario 2 by doubling the volume of reused water. Scenario 4 extends scenario 3 by also doubling the volume of desalinated water. Scenario 5 combines the 10% improvement in irrigation efficiency from scenario 1 with a doubling of both reused and desalinated water volumes, along with a stabilization of irrigated areas. Scenario 6 adds a 7% increase in water supply to the measures in scenario 5. Finally, scenario 7 combines the 10% irrigation efficiency improvement from scenario 1 with a doubling of reused and desalinated water volumes, but reduces the irrigated area by 15%. This study is of crucial importance to decision-makers, as it provides them with strategies for promoting water-saving practices and, consequently, advancing the sustainable development agenda. © 2025 The Authors

关键词： Climate change Integrated water management Scenario System dynamics

来源：评论

学校读者我要写书评

暂无评论

A Survey on Graph Neural Networks and Graph Transformers in computer vision: A Task-Oriented Perspective

arXiv

引用

arXiv 2022年

作者： Chen, Chaoqi Wu, Yushuang Dai, Qiyuan Zhou, Hong-Yu Xu, Mutian Yang, Sibei Han, Xiaoguang Yu, Yizhou Department of Computer Science The University of Hong Kong Hong Kong School of Science and Engineering The Chinese University of Hong Kong Shenzhen China Future Network of Intelligence Institute CUHK-Shenzhen China School of Information Science and Technology ShanghaiTech University Shanghai China Shanghai Engineering Research Center of Intelligent Vision and Imaging China

Graph Neural Networks (GNNs) have gained momentum in graph representation learning and boosted the state of the art in a variety of areas, such as data mining (e.g., social network analysis and recommender systems), computer vision (e.g., object detection and point cloud learning), and natural language processing (e.g., relation extraction and sequence learning), to name a few. With the emergence of Transformers in natural language processing and computer vision, graph Transformers embed a graph structure into the Transformer architecture to overcome the limitations of local neighborhood aggregation while avoiding strict structural inductive biases. In this paper, we present a comprehensive review of GNNs and graph Transformers in computer vision from a task-oriented perspective. Specifically, we divide their applications in computer vision into five categories according to the modality of input data, i.e., 2D natural images, videos, 3D data, vision + language, and medical images. In each category, we further divide the applications according to a set of vision tasks. Such a task-oriented taxonomy allows us to examine how each task is tackled by different GNN-based approaches and how well these approaches perform. Based on the necessary preliminaries, we provide the definitions and challenges of the tasks, in-depth coverage of the representative approaches, as well as discussions regarding insights, limitations, and future directions. Copyright © 2022, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

EdgeML: Towards Network-Accelerated Federated Learning over Wireless Edge

arXiv

引用

arXiv 2021年

作者： Pinyoanuntapong, Pinyarash Janakaraj, Prabhu Balakrishnan, Ravikumar Lee, Minwoo Chen, Chen Wang, Pu The Department of Computer Science University of North Carolina at Charlotte CharlotteNC28223 United States Intel Labs United States The Center for Research in Computer Vision University of Central Florida OrlandoFL32816 United States

Federated learning (FL) is a distributed machine learning technology for next-generation AI systems that allows a number of workers, i.e., edge devices, collaboratively learn a shared global model while keeping their data locally to prevent privacy leakage. Enabling FL over wireless multi-hop networks can democratize AI and make it accessible in a costeffective manner. However, the noisy bandwidth-limited multihop wireless connections can lead to delayed and nomadic model updates, which significantly slows down the FL convergence speed. To address such challenges, this paper aims to accelerate FL convergence over wireless edge by optimizing the multi-hop federated networking performance. In particular, the FL convergence optimization problem is formulated as a Markov decision process (MDP). To solve such MDP, multiagent reinforcement learning (MA-RL) algorithms along with domain-specific action space refining schemes are developed, which online learn the delay-minimum forwarding paths to minimize the model exchange latency between the edge devices (i.e., workers) and the remote server. To validate the proposed solutions, FedEdge is developed and implemented, which is the first experimental framework in the literature for FL over multihop wireless edge computing networks. FedEdge allows us to fast prototype, deploy, and evaluate novel FL algorithms along with RL-based system optimization methods in real wireless devices. Moreover, a physical experimental testbed is implemented by customizing the widely adopted Linux wireless routers and ML computing nodes. Such testbed can provide valuable insights into the practical performance of FL in the field. Finally, our experimentation results on the testbed show that the proposed networkaccelerated FL system can practically and significantly improve FL convergence speed, compared to the FL system empowered by the production-grade commercially-available wireless networking protocol, BATMAN-Adv. © 2021, CC BY.

关键词： Testbeds

来源：评论

学校读者我要写书评

暂无评论

Tunable unidirectional surface plasmon-polaritons at the interface between gyrotropic and isotropic conductors 22

Tunable unidirectional surface plasmon-polaritons at the int...

引用

22nd International Conference on Electromagnetics in Advanced Applications, ICEAA 2021

作者： Liang, Y. Pakniyat, S. Xiang, Y. Chen, J. Shi, F. Hanson, G.W. Cen, C. West Virginia University Department of Physics and Astronomy MorgantownWV26506 United States University of Wisconsin-Milwaukee Department of Electrical Engineering MilwaukeeWI53211 United States University of Pittsburgh Peterson Institute of Nano Science and Engineering Department of Electrical and Computer Engineering PittsburghPA15261 United States Guangxi Key Lab for Relativistic Astrophysics Center on Nanoenergy Research School of Physical Science and Technology Guangxi University Nanning Guangxi530004 China Key Laboratory of Computer Vision and System of Ministry of Education Tianjin University of Technology Tianjin300384 China

ISBN: (纸本)9781665413862

The propagation of SPPs in solid state systems is usually reciprocal, which means that a forward-traveling SPP wave can be easily backscattered by disorders. Since such backscattering is detrimental for SPP-based information processing applications, the development of nonreciprocal plasmonic platforms that enforce unidirectional SPP propagations is of great importance. © 2021 IEEE.

关键词： Electromagnetic wave polarization

来源：评论

学校读者我要写书评

暂无评论

Heterogeneous Oblique Double Random Forest

arXiv

引用

arXiv 2023年

作者： Ganaie, M.A. Tanveer, M. Beheshti, I. Ahmad, N. Suganthan, P.N. Department of Robotics University of Michigan Ann ArborMI United States Department of Mathematics Indian Institute of Technology Indore Simrol Indore453552 India Department of Human Anatomy and Cell Science Rady Faculty of Health Sciences Max Rady College of Medicine University of Manitoba WinnipegMB Canada Department of International Program of Electrical Engineering and Computer Science National Taipei University of Technology Taipei Taiwan KINDI Center for Computing Research College of Engineering Qatar University Doha Qatar

The decision tree ensembles use a single data feature at each node for splitting the data. However, splitting in this manner may fail to capture the geometric properties of the data. Thus, oblique decision trees generate the oblique hyperplane for splitting the data at each non-leaf node. Oblique decision trees capture the geometric properties of the data and hence, show better generalization. The performance of the oblique decision trees depends on the way oblique hyperplanes are generate and the data used for the generation of those hyperplanes. Recently, multiple classifiers have been used in a heterogeneous random forest (RaF) classifier, however, it fails to generate the trees of proper depth. Moreover, double RaF studies highlighted that larger trees can be generated via bootstrapping the data at each non-leaf node and splitting the original data instead of the bootstrapped data recently. The study of heterogeneous RaF lacks the generation of larger trees while as the double RaF based model fails to take over the geometric characteristics of the data. To address these shortcomings, we propose heterogeneous oblique double RaF. The proposed model employs several linear classifiers at each non-leaf node on the bootstrapped data and splits the original data based on the optimal linear classifier. The optimal hyperplane corresponds to the models based on the optimized impurity criterion. The experimental analysis indicates that the performance of the introduced heterogeneous double random forest is comparatively better than the baseline models. To demonstrate the effectiveness of the proposed heterogeneous double random forest, we used it for the diagnosis of Schizophrenia disease. The proposed model predicted the disease more accurately compared to the baseline models. © 2023, CC BY-NC-ND.

关键词： Decision trees

来源：评论

学校读者我要写书评

暂无评论

Enhancing UAV Autonomous Navigation in Indoor Environments Using Reinforcement Learning and Convolutional Neural Networks

Enhancing UAV Autonomous Navigation in Indoor Environments U...

引用

International Symposium on Intelligent Systems and Informatics (SISY)

作者： Homayoun Safarpour Motealegh Mahalegi Amirfarhad Farhadi György Molnár Eniko Nagy Obuda University Institute of Mechatronics and Vehicle Engineering Budapest Hungary Department of Computer Engineering Science and Research Branch Islamic Azad University Tehran Iran Apaczai Csere Janos Faculty of Humanities Education and Social Sciences Kando Kalman Faculty of Electrical Engineering Budapest Obuda University Hungary Szechenyi University Gyor Hungary Antal Bejczy Center for Intelligent Robotics John von Neumann Faculty of Informatics Obuda University Budapest Hungary

ISBN: (数字)9798350385601

ISBN: (纸本)9798350385618

In recent years, the number of unmanned aerial vehicle (UAV) applications has increased. However, navigating them indoors is still tricky because no GPS signals are available, and the obstacles constantly change. This study exploits the collaboration between convolutional neural networks (CNNs) and reinforcement learning (RL) to overcome these challenges, improving the level of independence of UAVs in indoor settings. We present a new technique known as potential-based reward shaping, which directly incorporates expert knowledge into the reinforcement learning framework. This technique efficiently addresses the problem of sparse rewards that often hinder learning in conventional RL configurations. In addition, we have created a reflective reinforcement learning agent that systematically assesses its actions to prioritize those that result in the most substantial enhancements. We have extensively evaluated the performance of our method with the High-Definition Indoor Navigation (HDIN) dataset to show that our method continuously outperformed current non-introspective RL techniques in terms of decision-making speed and navigation accuracy in this evaluation. The results show that this new CNN-RL combination is feasible and suggests a scalable method for complex indoor navigation problems.

关键词： Indoor navigation Refining Reinforcement learning Autonomous aerial vehicles Indoor environment Convolutional neural networks Intelligent systems Informatics Autonomous robots Global Positioning System

来源：评论

学校读者我要写书评

暂无评论

Gesture Control Method for Oral Blood Mixed Liquid Removal Robot∗ 5

Gesture Control Method for Oral Blood Mixed Liquid Removal R...

引用

5th World Robot Conference Symposium on Advanced robotics and Automation, WRC SARA 2023

作者： Ma, Xudong Kuang, Ziao Li, Han Chen, Anqi Li, Changsheng Liu, Wenyong Wang, Junchen Hu, Yida David Gao, Li Su, Bai-Quan Beijing University of Posts and Tele-communications Medical Robotics Laboratory School of Automation Beijing100876 China Beijing University of Posts and Tele-communications School of Computer Beijing100876 China Beijing Institute of Technology School of Mechatronical Engineering Beijing100081 China Beihang University Key Laboratory for Biomechanics and Mechanobiology of Chinese Education Ministry Beijing Advanced Innovation Centre for Biomedical Engineering School of Biological Science and Medical Engineering Beijing100083 China Beihang University School of Mechanical Engineering and Automation Beijing100083 China Brigham and Women's Hospital Harvard Medical School BostonMA02115 United States Peking University School and Hospital of Stomatology National Clinical Research Center for Oral Diseases National Stomatological Center Department of Periodontology Beijing China

ISBN: (纸本)9798350307320

All of oral surgeries are accompanied by bleeding. Manual removal of blood is carried out by a nurse and dozens of repetitive blood suction operations are implemented by the nurse during a single oral surgery. In our previous work, a robot for removal of blood-water mixed liquid is designed. However, the robot requires an initial order to autonomous removal of blood mixed liquid. In this paper, a hand gesture control method for typical operations of the robot is developed. Gesture dataset are developed at first and a deep learning neural network is trained for detection of hand gesture. The gesture control procedures for four typical operations are provided in detail. In a group of experiments, it's demonstrated that an oral surgeon controls the robot's motion by presenting prescribed hand gesture. The proposed method could be an alternative manner for blood mixed liquid removal. © 2023 IEEE.

关键词： Blood

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：