检索结果-内蒙古大学图书馆

IEEE International Conference on e-Science and Grid computing

作者： Zilinghan Li Shilan He Pranshu Chaturvedi Trung-Hieu Hoang Minseok Ryu E. A. Huerta Volodymyr Kindratenko Jordan Fuhrman Maryellen Giger Ryan Chard Kibaek Kim Ravi Madduri Data Science and Learning Division Argonne National Laboratory Lemont IL USA Department of Computer Science University of Illinois at Urbana-Champaign Urbana IL USA National Center for Supercomputing Applications University of Illinois at Urbana-Champaign Urbana IL USA Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign Urbana IL USA School of Computing and Augmented Intelligence Arizona State University Tempe AZ USA Department of Physics University of Illinois at Urbana-Champaign Urbana IL USA Department of Computer Science The University of Chicago Chicago IL USA Mathematics and Computer Science Division Argonne National Laboratory Lemont IL USA University of Chicago Consortium for Advanced Science and Engineering Chicago IL USA

Cross-silo privacy-preserving federated learning (PPFL) is a powerful tool to collaboratively train robust and generalized machine learning (ML) models without sharing sensitive (e.g., healthcare of financial) local data. To ease and accelerate the adoption of PPFL, we introduce APPFLx, a ready-to-use platform that provides privacy-preserving cross-silo federated learning as a service. APPFLx employs Globus authentication to allow users to easily and securely invite trustworthy collaborators for PPFL, implements several synchronous and asynchronous FL algorithms, streamlines the FL experiment launch process, and enables tracking and visualizing the life cycle of FL experiments, allowing domain experts and ML practitioners to easily orchestrate and evaluate cross-silo FL under one platform. APPFLx is available online at https://***

关键词：

来源：评论

学校读者我要写书评

暂无评论

PUGAN: Physical Model-Guided Underwater Image Enhancement Using GAN with Dual-Discriminators

arXiv

引用

arXiv 2023年

作者： Cong, Runmin Yang, Wenyu Zhang, Wei Li, Chongyi Guo, Chun-Le Huang, Qingming Kwong, Sam Institute of Information Science Beijing Jiaotong University Beijing100044 China School of Control Science and Engineering Shandong University Jinan250061 China Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan250061 China Beijing Key Laboratory of Advanced Information Science and Network Technology Beijing100044 China College of Computer Science Nankai University Tianjin300350 China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing101408 China Key Laboratory of Intelligent Information Processing Institute of Computing Technology Chinese Academy of Sciences Beijing100190 China Peng Cheng Laboratory Shenzhen518055 China Department of Computer Science City University of Hong Kong Hong Kong City University of Hong Kong Shenzhen Research Institute Shenzhen51800 China

Due to the light absorption and scattering induced by the water medium, underwater images usually suffer from some degradation problems, such as low contrast, color distortion, and blurring details, which aggravate the difficulty of downstream underwater understanding tasks. Therefore, how to obtain clear and visually pleasant images has become a common concern of people, and the task of underwater image enhancement (UIE) has also emerged as the times require. Among existing UIE methods, Generative Adversarial Networks (GANs) based methods perform well in visual aesthetics, while the physical model-based methods have better scene adaptability. Inheriting the advantages of the above two types of models, we propose a physical model-guided GAN model for UIE in this paper, referred to as PUGAN. The entire network is under the GAN architecture. On the one hand, we design a Parameters Estimation subnetwork (Par-subnet) to learn the parameters for physical model inversion, and use the generated color enhancement image as auxiliary information for the Two-Stream Interaction Enhancement subnetwork (TSIE-subnet). Meanwhile, we design a Degradation Quantization (DQ) module in TSIE-subnet to quantize scene degradation, thereby achieving reinforcing enhancement of key regions. On the other hand, we design the Dual-Discriminators for the style-content adversarial constraint, promoting the authenticity and visual aesthetics of the results. Extensive experiments on three benchmark datasets demonstrate that our PUGAN outperforms state-of-the-art methods in both qualitative and quantitative metrics. The code and results can be found from the link of https://***/proj_***. © 2023, CC BY-NC-SA.

关键词： Discriminators

来源：评论

学校读者我要写书评

暂无评论

PuzzleTuning: Explicitly Bridge Pathological and Natural Image with Puzzles

arXiv

引用

arXiv 2023年

作者： Zhang, Tianyi Lyu, Shangqing Lei, Yanli Chen, Sicheng Ying, Nan He, Yufang Liu, Zeyu Zhao, Yu Feng, Yunlu Jin, Yueming Lee, Hwee Kuan Zhang, Guanglei The Beijing Advanced Innovation Center for Biomedical Engineering School of Biological Science and Medical Engineering Beihang University Beijing100191 China Singapore138410 Singapore The School of Microelectronics Xi’an Jiaotong University Xi’an710049 China The Department of Pathology Peking Union Medical College Hospital Beijing100006 China The Department of Gastroenterology Peking Union Medical College Hospital Beijing100006 China The Department of Biomedical Engineering National University of Singapore Singapore117417 Singapore Department of Electrical and Computer Engineering National University of Singapore Singapore117417 Singapore The School of Computing National University of Singapore Singapore117417 Singapore School of Biological Sciences Nanyang Technological University 639798 Singapore 169856 Singapore International Research Laboratory on Artifical Intelligence 138632 Singapore Singapore Institute for Clinical Sciences Singapore117609 Singapore

Pathological image analysis is a crucial field in computer vision. Due to the annotation scarcity in the pathological field, pre-training with self-supervised learning (SSL) is widely applied to learn on unlabeled images. However, the current SSL-based pathological pre-training: (1) does not explicitly explore the essential focuses of the pathological field, and (2) does not effectively bridge with and thus take advantage of the knowledge from natural images. To explicitly address them, we propose our large-scale PuzzleTuning framework, containing the following innovations. Firstly, we define three task focuses that can effectively bridge knowledge of pathological and natural domain: appearance consistency, spatial consistency, and restoration understanding. Secondly, we devise a novel multiple puzzle restoring task, which explicitly pre-trains the model regarding these focuses. Thirdly, we introduce an explicit prompt-tuning process to incrementally integrate the domain-specific knowledge. It builds a bridge to align the large domain gap between natural and pathological images. Additionally, a curriculum-learning training strategy is designed to regulate task difficulty, making the model adaptive to the puzzle restoring complexity. Experimental results show that our PuzzleTuning framework outperforms the previous state-of-the-art methods in various downstream tasks on multiple datasets. The code, demo, and pre-trained weights are available at https://***/sagizty/PuzzleTuning. © 2023, CC BY-NC-SA.

关键词： Image analysis

来源：评论

学校读者我要写书评

暂无评论

A novel hyperchaotic multistable heterogeneous neural network with hidden attractors and its application in image encryption

引用

Chinese Journal of Physics 2025年

作者： Zhi Huang Zhen Li Qiao Wang Weijie Tan Xianming Wu College of Big Data and Information Engineering Guizhou University Guiyang 550025 China State Key Laboratory of Public Big Data Guizhou University Guiyang 550025 China Guizhou Key Laboratory of Artificial Intelligence and Brain-inspired Computing College of Mathematics and Big Data Guizhou Education University Guiyang 550018 China State Key Laboratory of Public Big Data Key Laboratory of Advanced Manufacturing Technology of Ministry of Education Guizhou University Guiyang 550025 China School of Mechanical and Electrical Engineering Guizhou Normal University Guiyang 550025 China

By simulating the unique properties of different types of neurons and their interconnecting patterns, heterogeneous neural networks can more accurately reflect the structural and functional characteristics of biological neural systems. Therefore, a novel hyperbolic non-volatile locally active memristor is proposed, enabling the construction of a heterogeneous Hindmarsh-Rose neuron memristive synapse-coupled Hopfield neural network (HR-M-HNN) via synaptic characteristic emulation. Dynamic analysis reveals that the HR-M-HNN exhibits hidden attractor characteristics due to the absence of equilibrium points and can generate complex dynamics behaviors, including hyperchaos, chaos, periodic and quasi-periodic, and multistability. Further research reveals a strong correlation between the chaotic state of HR-M-HNN and the locally active control parameter of the memristor. To verify the accuracy of the dynamic analysis, the corresponding analog circuit is designed. Furthermore, based on the HR-M-HNN hyperchaotic system, a novel image encryption scheme is proposed, which utilizes the position index of the Latin square to perform synchronous permutation and diffusion on the plain image, reducing redundant operations and effectively enhancing the encryption speed. Security analyses demonstrate that the proposed image encryption scheme offers excellent performance and robustness.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A novel extreme multistability system coupled by HR neuron and unidirectional cyclic HNN and its application in medical image encryption

引用

Chaos, Solitons & Fractals 2025年 199卷

In brain-like research, investigating the impact of heterogeneous neurons on specific neural network structures is of significant importance. In this paper, a heterogeneous memristive synapse-coupled system (HMSCS) is proposed, which is coupled by Hindmarsh–Rose (HR) and unidirectional cyclic Hopfield neural network (HNN), and is established by presenting a new locally active memristor with multi-stable behavior to simulate the synaptic connection between HR neuron and HNN. The dynamic analysis results show that the locally active region of the memristor can significantly affect the chaos range of HMSCS, and the initial value of the memristor can drive a shift in the phase of the attractor, resulting in extreme multistability. In addition, the numerical analyses are validated by circuit design and simulation. Furthermore, a novel medical image encryption scheme is proposed based on the HMSCS chaotic system, in which an improved Hilbert curve permutation method is given, where the traversal direction and starting point of the Hilbert curve are dynamically selected in each round, thereby effectively enhance permutation performance. Additionally, a region-constrained initial value mapping method is proposed to prevent multistability systems from entering periodic orbits due to improper initial value selection, which is a common problem in existing multistability chaotic systems applied to image encryption. Some common security analyses results show that our proposed encryption scheme has excellent security performance for medical image application scenarios.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Graphene quantum dots induced performance enhancement in memristors

引用

Nanoscale 2025年第23期17卷 2025 May 28页

作者： Jintao He Guangdong Zhou Bai Sun Lingpeng Yan Xiaochen Lang Yongzhen Yang Haotian Hao MOE Key Laboratory of Interface Science and Engineering in Advanced Materials College of Materials Science and Engineering Taiyuan University of Technology Taiyuan 030024 China. yangyongzhen@***. College of Artificial Intelligence Brain-Inspired Computing & Intelligent Control of Chongqing Key Lab Southwest University Chongqing 400715 China. Frontier Institute of Science and Technology (FIST) Xi'an Jiaotong University Xi'an Shaanxi 710049 China. College of Artificial Intelligence Institute of Biomedical Engineering Taiyuan University of Technology Taiyuan 030024 China. haohaotian@***.

With the rapid development of information technology, the demand for miniaturization, integration, and intelligence of electronic devices is growing rapidly. As a key device in the non-von Neumann architecture, memristors can perform computations while storing data, enhancing computational efficiency and reducing power consumption. Memristors have become pivotal in driving the advancement of artificial intelligence (AI) and Internet of Things technologies. Combining the electronic properties of graphene with the size effects of quantum dots, graphene quantum dot (GQD)-based memristors exhibit potential applications in constructing brain-inspired neuromorphic computing systems and achieving AI hardware acceleration, making them a focal point of research interest. This review provides an overview of the preparation, mechanism, and application of GQD-based memristors. Initially, the structure, properties, and synthesis methods of GQDs are introduced in detail. Subsequently, the memristive mechanisms of GQD-based memristors are presented from three perspectives: the metal conductive filament mechanism, the electron trapping and detrapping mechanism, and the oxygen vacancy conductive filament mechanism. Furthermore, the different application scenarios of GQD-based memristors in both digital and analog types are summarized, encompassing information storage, brain-like artificial synapses, visual perception systems, and brain-machine interfaces. Finally, the challenges and future development prospects of GQD-based memristors are discussed.

关键词： Quantum computers

来源：评论

学校读者我要写书评

暂无评论

Neural-Network-Based Control of Wheeled Mobile Manipulators With Unknown Kinematic Models

Neural-Network-Based Control of Wheeled Mobile Manipulators ...

引用

International Symposium on Autonomous Systems (ISAS)

作者： Ning Tan Zelei Zhu Peng Yu School of Computer Science and Engineering Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Sun Yat-sen University Guangzhou China School of Computer Science and Engineering Sun Yat-sen University Guangzhou China

ISBN: (数字)9781665418829

ISBN: (纸本)9781665447997

As we all known, tracking control is always a basic and important issue in robot industry. In the past, the method based on the pseudo-inverse of Jacobian matrix is a conventional solution which has two major limitations: one is the position error accumulation phenomenon; the other is the requirement for the model information of robots. Therefore, in this paper, we propose a neural-network-based control (NNBC) scheme for wheeled mobile manipulators. The NNBC method overcomes the two limitations and is proved feasible to control the track of stationary robot manipulators. However, in robot industry, mobile manipulators controlling is an unavoidable topic. It is necessary to confirm the NNBC method is workable for mobile manipulators. Thus in this paper, we are going to discuss how to control the track of wheeled mobile manipulators using the NNBC method. In details, we will control the joints of the stationary manipulator and the wheels of the mobile platform to ensure the end-effector make a correct track without knowing the information of the parameters and structure of wheeled mobile manipulators. By integrating the stationary manipulator and the mobile platform into a system, the wheels and the manipulator can cooperate correctly to complete the task of the end-effector. Finally, some simulations will be performed to prove that the NNBC method is workable for wheeled mobile manipulators.

关键词： Industries Service robots Wheels Artificial neural networks Kinematics Manipulators Mobile robots

来源：评论

学校读者我要写书评

暂无评论

Detection of Deepfake videos using long distance attention

arXiv

引用

arXiv 2021年

作者： Lu, Wei Liu, Lingyi Luo, Junwei Zhao, Xianfeng Zhou, Yicong Huang, Jiwu School of Computer Science and Engineering Guangdong Province Key Laboratory of Information Security Technology Ministry of Education Key Laboratory of Machine Intelligence and Advanced Computing Sun Yat-sen University Guangzhou510006 China State Key Laboratory of Information Security Institute of Information Engineering Chinese Academy of Sciences Beijing100195 China School of Cyber Security University of Chinese Academy of Sciences Beijing100195 China Department of Computer and Information Science University of Macau 999078 China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen Key Laboratory of Media Security Shenzhen University Shenzhen518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen518055 China

With the rapid progress of deepfake techniques in recent years, facial video forgery can generate highly deceptive video contents and bring severe security threats. And detection of such forgery videos is much more urgent and challenging. Most existing detection methods treat the problem as a vanilla binary classification problem. In this paper, the problem is treated as a special fine-grained classification problem since the differences between fake and real faces are very subtle. It is observed that most existing face forgery methods left some common artifacts in the spatial domain and time domain, including generative defects in the spatial domain and inter-frame inconsistencies in the time domain. And a spatial-temporal model is proposed which has two components for capturing spatial and temporal forgery traces in global perspective respectively. The two components are designed using a novel long distance attention mechanism. The one component of the spatial domain is used to capture artifacts in a single frame, and the other component of the time domain is used to capture artifacts in consecutive frames. They generate attention maps in the form of patches. The attention method has a broader vision which contributes to better assembling global information and extracting local statistic information. Finally, the attention maps are used to guide the network to focus on pivotal parts of the face, just like other fine-grained classification methods. The experimental results on different public datasets demonstrate that the proposed method achieves the state-of-the-art performance, and the proposed long distance attention method can effectively capture pivotal parts for face forgery. Copyright © 2021, The Authors. All rights reserved.

关键词： Time domain analysis

来源：评论

学校读者我要写书评

暂无评论

Controlling Robot Manipulators Using Gradient-Based Recursive Neural Networks

Controlling Robot Manipulators Using Gradient-Based Recursiv...

引用

International Symposium on Autonomous Systems (ISAS)

作者： Ning Tan Mao Zhang Peng Yu School of Computer Science and Engineering Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Sun Yat-sen University Guangzhou China School of Computer Science and Engineering Sun Yat-sen University Guangzhou China

ISBN: (数字)9781665418829

ISBN: (纸本)9781665447997

Discrete robot manipulators have been a simple combination of several serially connected links and actuated joints. In addition, these manipulators have total degrees of freedom and prove to perform well in many fields. The motion of traditional robot is generated through the movement in discrete position, i.e., joints. A gradient neural network control is proposed for approximating the system uncertainty, such as the tracking of robot manipulators with model-free shown in this paper. Based on the method proposed above, the simulation results testify the validity and accuracy of our models.

关键词： Uncertainty Tracking Simulation Neural networks Manipulators Trajectory Task analysis

来源：评论

学校读者我要写书评

暂无评论

Model-Less Feedback Control for Soft Manipulators with Jacobian Adaptation

Model-Less Feedback Control for Soft Manipulators with Jacob...

引用

International Symposium on Autonomous Systems (ISAS)

作者： Yu-Yang Wu Ning Tan School of Computer Science and Engineering Sun Yat-sen University Guangzhou China School of Computer Science and Engineering Key Laboratory of Machine Intelligence and Advanced Computing Ministry of Education Sun Yat-sen University Guangzhou China

ISBN: (数字)9781665418829

ISBN: (纸本)9781665447997

In this paper we propose a novel closed-loop model-less feedback control scheme for soft manipulators with Jacobian matrix adaptation. As the major merit, this approach does not rely on a specific kinematic model and thus has wide applicability and portability for different kinds of soft manipulators. We make use of closed-loop control, convex optimization and adaptation strategy to estimate the Jacobian matrix for trajectory tracking tasks. In the simulation validation, we adapt the algorithm to a two-section soft manipulator of piece-wise constant curvature to track a desired trajectory. Furthermore, we compare our algorithm with two other model-less algorithms and a model-based algorithm with noise on their tracking performances. Validation results prove the feasibility of this simple yet effective model-less feedback control scheme.

关键词： Jacobian matrices Adaptation models Trajectory tracking Manipulators Feedback control Trajectory Task analysis

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：