检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Lv, Fengling Shang, Xinyi Zhou, Yang Zhang, Yiqun Li, Mengke Lu, Yang Key Laboratory of Multimedia Trusted Perception and Efficient Computing Xiamen University China University College London United Kingdom A*STAR Institute of High Performance Computing Singapore School of Computer Science and Technology Guangdong University of Technology Guangzhou China Guangdong Laboratory of Artificial Intelligence and Digital Economy Shenzhen China Fujian Key Laboratory of Sensing and Computing for Smart City School of Informatics Xiamen University China

Personalized Federated Learning (PFL) aims to acquire customized models for each client without disclosing raw data by leveraging the collective knowledge of distributed clients. However, the data collected in real-world scenarios is likely to follow a long-tailed distribution. For example, in the medical domain, it is more common for the number of general health notes to be much larger than those specifically related to certain diseases. The presence of long-tailed data can significantly degrade the performance of PFL models. Additionally, due to the diverse environments in which each client operates, data heterogeneity is also a classic challenge in federated learning. In this paper, we explore the joint problem of global long-tailed distribution and data heterogeneity in PFL and propose a method called Expert Collaborative Learning (ECL) to tackle this problem. Specifically, each client has multiple experts, and each expert has a different training subset, which ensures that each class, especially the minority classes, receives sufficient training. Multiple experts collaborate synergistically to produce the final prediction output. Without special bells and whistles, the vanilla ECL outperforms other state-of-the-art PFL methods on several benchmark datasets under different degrees of data heterogeneity and long-tailed distribution. Copyright © 2024, The Authors. All rights reserved.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Deep Neural Cellular Potts Models

arXiv

引用

arXiv 2025年

作者： Minartz, Koen d’Hondt, Tim Hillmann, Leon Starruß, Jörn Brusch, Lutz Menkovski, Vlado Department of Mathematics and Computer Science Eindhoven University of Technology Eindhoven Netherlands Department of Applied Physics and Science Education Eindhoven University of Technology Eindhoven Netherlands Department of Biomedical Engineering Eindhoven University of Technology Eindhoven Netherlands Center for Information Services and High Performance Computing TUD Dresden University of Technology Dresden Germany

The cellular Potts model (CPM) is a powerful computational method for simulating collective spatiotemporal dynamics of biological cells. To drive the dynamics, CPMs rely on physics-inspired Hamiltonians. However, as first principles remain elusive in biology, these Hamiltonians only approximate the full complexity of real multicellular systems. To address this limitation, we propose NeuralCPM, a more expressive cellular Potts model that can be trained directly on observational data. At the core of NeuralCPM lies the Neural Hamiltonian, a neural network architecture that respects universal symmetries in collective cellular dynamics. Moreover, this approach enables seamless integration of domain knowledge by combining known biological mechanisms and the expressive Neural Hamiltonian into a hybrid model. Our evaluation with synthetic and real-world multicellular systems demonstrates that NeuralCPM is able to model cellular dynamics that cannot be accounted for by traditional analytical Hamiltonians. Copyright © 2025, The Authors. All rights reserved.

关键词： Potts model

来源：评论

学校读者我要写书评

暂无评论

Intensity correlations in measurement-device-independent quantum key distribution

arXiv

引用

arXiv 2024年

作者： Liu, Junxuan Xing, Tianyi Liu, Ruiyin Chen, Zihao Tan, Hao Huang, Anqi Institute for Quantum Information State Key Laboratory of High Performance Computing College of Computer Science and Technology National University of Defense Technology Changsha410073 China China Telecom Quantum Information Technology Group Co. Ltd Hefei230088 China

The intensity correlations due to imperfect modulation during the quantum-state preparation in a measurement-device-independent quantum key distribution (MDI QKD) system compromise its security performance. Therefore, it is crucial to assess the impact of intensity correlations on the practical security of MDI QKD systems. In this work, we propose a theoretical model that quantitatively analyzes the secure key rate of MDI QKD systems under intensity correlations. Furthermore, we apply the theoretical model to a practical MDI QKD system with measured intensity correlations, which shows that the system struggles to generate keys efficiently under this model. We also explore the boundary conditions of intensity correlations to generate secret keys. This study extends the security analysis of intensity correlations to MDI QKD protocols, providing a methodology to evaluate the practical security of MDI QKD systems. Copyright © 2024, The Authors. All rights reserved.

关键词： Quantum cryptography

来源：评论

学校读者我要写书评

暂无评论

Generative Methods for Out-of-distribution Prediction and Applications for Threat Detection and Analysis: A Short Review 1st

Generative Methods for Out-of-distribution Prediction and Ap...

引用

1st International Workshop on Digital Sovereignty in Cyber Security: New Challenges in Future Vision, CyberSec4Europe 2022

作者： Coppolillo, Erica Liguori, Angelica Guarascio, Massimo Pisani, Francesco Sergio Manco, Giuseppe Department of Mathematics and Computer Science University of Calabria Rende Italy Department of Computer Engineering Modeling Electronics an d Systems University of Calabria Rende Italy Institute for High Performance Computing and Networking of Italian National Research Council Rende Italy

ISBN: (纸本)9783031360954

In recent times, Machine Learning has played an important role in developing novel advanced tools for threat detection and mitigation. Intrusion Detection, Misinformation, Malware, and Fraud Detection are just some examples of cybersecurity fields in which Machine Learning techniques are used to reveal the presence of malicious behaviors. However, Out-of-Distribution, i.e., the potential distribution gap between training and test set, can heavily affect the performances of the traditional Machine Learning based methods. Indeed, they could fail in identifying out-of-samples as possible threats, therefore devising robust approaches to cope with this issue is a crucial and relevant challenge to mitigate the risk of undetected attacks. Moreover, a recent emerging line proposes to use generative models to yield synthetic likely examples to feed the learning algorithms. In this work, we first survey recent Machine Learning and Deep Learning based solutions to face both the problems, i.e., outlier detection and generation;then we illustrate the main cybersecurity application scenarios in which these approaches have been adopted successfully. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

Sclabel: Finding performance Bugs Through Clone Detection Based on Labeled Code Templates

SSRN

引用

SSRN 2022年

作者： Deng, Zihan Wang, Wenjie Xue, Yinxing Xu, Yun School of Computer Science and Technology University of Science and Technology of China Hefei230027 China Key Laboratory on High Performance Computing of Anhui Province Hefei230027 China

performance bugs are defects in source code that slow down program execution, and produced causes of them are complex and diverse. Many tools have been proposed for detecting performance bugs, but they can only find certain types of results which having similar code patterns, and need dynamic analysis or control flow graph of the program. Therefore, conventional tools are lack of generality and high cost of space and time. As we know, many state-of-the-art clone detection techniques have been proposed for general similar code detection, but they can only detect highly similar code (e.g., tools based on text, token or AST) or rely on training dataset (e.g., tools based on machine learning), making them impractical for detecting performance bugs in real-world projects. © 2022, The Authors. All rights reserved.

关键词： Flow graphs

来源：评论

学校读者我要写书评

暂无评论

REFROM: Responsive, Energy-Efficient Frame Rendering for Mobile Devices

REFROM: Responsive, Energy-Efficient Frame Rendering for Mob...

引用

International Symposium on Low Power Electronics and Design (ISLPED)

作者： Tsung-Yen Hsu Yi-Shen Chen Yun-Chih Chen Yuan-Hao Chang Tei-Wei Kuo Department of Computer Science and Information Engineering National Taiwan University Department of Electronic and Computer Engineering National Taiwan University of Science and Technology Institute of Information Science Academia Sinica Mohamed bin Zayed University of Artificial Intelligence High Performance and Scientific Computing Center National Taiwan University

The increasing demand for high-quality graphics on mobile devices necessitates a high frame rate for display refresh. However, current process scheduling and memory management policies fail to consider the computation demands of frame rendering because they are optimized for saving energy and resource utilization. This leads to unresponsive displays for mobile users due to rendering delays. Accurately estimating computation demands is challenging for the mobile operating system, particularly under memory pressure, without display-specific semantics from user space. Moreover, the complexity of frame rendering makes it infeasible to schedule them with real-time policies. To address these issues, we propose a new framework called REFROM that utilizes a history-based frame time estimator to analyze frame time samples from UI threads and predict the computation requirements of upcoming frames. Experimental results demonstrate that REFROM reduces the number of delayed frames by up to 40% and improves up to 4% energy efficiency compared to the existing approaches.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Critical hybridization of skin modes in coupled non-Hermitian chains

引用

Physical Review Research 2022年第1期4卷 013243-013243页

作者： S. M. Rafi-Ul-Islam Zhuo Bin Siu Haydar Sahin Ching Hua Lee Mansoor B. A. Jalil Department of Electrical and Computer Engineering National University of Singapore Singapore Institute of High Performance Computing Agency for Science Technology and Research Singapore Department of Physics National University of Singapore Singapore

Non-Hermitian topological systems exhibit a plethora of unusual topological phenomena including the extreme localization of eigenstates, known as the non-Hermitian skin effect (NHSE), which occurs in open chains. However, the non-Hermitian characteristics of coupled non-Hermitian chains remain largely unexplored. Here, we report on the eigenstate localization in coupled non-Hermitian chains with dissimilar inverse skin lengths in which the NHSE can be switched on and off by modulating the interchain coupling strength. In the limit of small interchain strength, the NHSE is present at both ends of the coupled system because of the weak hybridization of the eigenstates of the individual chains. The eigenspectrum under open boundary conditions (OBC) exhibits a discontinuous jump known as the critical NHSE (CNHSE) as the chain length increases. However, when the interchain coupling strength and hence, the hybridization between eigenstates become significant, the NHSE and CNHSE vanish. Instead, a peculiar “half-half skin localization” occurs in composite chains with opposite signs of inverse skin lengths, where half of the eigenstates are exponentially localized at one chain and the remainder of the eigenstates on the other chain. Our results provide new insights into the non-Hermitian phenomena in coupled systems.

关键词： Electronic structure Exotic phases of matter Skin effect Topological insulators Topological phase transition Topological phases of matter 1-dimensional systems Topological materials Band structure methods Bloch wave theory Lattice models in condensed matter Tight-binding model

来源：评论

学校读者我要写书评

暂无评论

Boundary-Preserving Parallel Refinement for Tetrahedral Meshes

Boundary-Preserving Parallel Refinement for Tetrahedral Mesh...

引用

high performance and Smart computing (HPSC), IEEE International Conference on

作者： Tinghao Yang Xiang Zhang Chunye Gong Xiang Gao Chuanfu Xu Jie Liu College of Computer Science and Technology National University of Defense Technology Changsha PR. China Laboratory of Digitizing Software for Frontier Equipment National University of Defense Technology Changsha PR. China State Key Laboratory of High Performance Computing National University of Defense Technology Changsha PR. China

ISBN: (数字)9798331596637

ISBN: (纸本)9798331596644

Mesh refinement is to increase the number of mesh elements by cutting each mesh element into multiple sub-ones. Though it can accelerate meshing and optimize mesh quality, the boundary of mesh elements as usual deviate from the respective complex geometry, especially with sharp curvatures. Besides, it demands more memory footprint for large problem size. Here, this paper devises a parallel boundary preserving mesh refinement method to addresses the above issues. Particularly, it seeks the longest edges for new point insertion so that mesh elements decouple with each other for high parallelism. More importantly, when the boudary mesh elements are refined, some new points sampled from the geometry will be inserted into their faces so that the new mesh elements are aware of boundary. As compared to refinement algorithms tailored for standalone machines and used by mainstream commercial softwares, the proposed refinement method offers promising efficiency.

关键词： Geometry Costs high performance computing Fitting Parallel processing Data structures Surface fitting Software Surface treatment Faces

来源：评论

学校读者我要写书评

暂无评论

Enabling All-to-Circular Polarization Up-Conversion by Nonlinear Chiral Metasurfaces with Rotational Symmetry

引用

Physical Review Letters 2025年第2期134卷 023804-023804页

作者： Dmitrii Gromyko Jun Siang Loh Jiangang Feng Cheng-Wei Qiu Lin Wu Science Mathematics and Technology (SMT) Singapore University of Technology and Design (SUTD) 8 Somapah Road 487372 Singapore Singapore Department of Electrical and Computer Engineering National University of Singapore 4 Engineering Drive 3 117583 Singapore Singapore Institute of High Performance Computing ASTAR (Agency for Science Technology and Research) 1 Fusionopolis Way No. 16-16 Connexis 138632 Singapore Singapore

We introduce a stacking strategy to design nonlinear chiral metasurfaces with high rotational symmetry, enabling degenerate quasi-bound-in-the-continuum resonances of absolute chirality. This symmetry allows for converting a circularly polarized pump into a circularly polarized nonlinear signal. Consequently, our rotation-symmetric bilayered metasurface design, tailored to respond solely to one specific circular polarization, can up-convert a linear or arbitrary polarized pump into circularly polarized nonlinear signals. The intensity ratios of these signals scale as the fourth (second) power of the chiral resonance amplitude for the second (third) harmonic.

关键词： Circular polarization

来源：评论

学校读者我要写书评

暂无评论

FundusGAN: A Hierarchical Feature-Aware Generative Framework for high-Fidelity Fundus Image Generation

arXiv

引用

arXiv 2025年

作者： Hou, Qingshan Wang, Meng Cao, Peng Ke, Zou Liu, Xiaoli Fu, Huazhu Zaiane, Osmar R. Computer Science and Engineering Northeastern University Shenyang China Key Laboratory of Intelligent Computing in Medical Image of Ministry of Education Northeastern University Shenyang China Alberta Machine Intelligence Institute University of Alberta Edmonton Canada Ophthalmology Yong Loo Lin School of Medicine National University of Singapore Singapore Institute of High Performance Computing Agency for Science Technology and Research Singapore

Recent advancements in ophthalmology foundation models such as RetFound have demonstrated remarkable diagnostic capabilities but require massive datasets for effective pre-training, creating significant barriers for development and deployment. To address this critical challenge, we propose FundusGAN, a novel hierarchical feature-aware generative framework specifically designed for high-fidelity fundus image synthesis. Our approach leverages a Feature Pyramid Network within its encoder to comprehensively extract multi-scale information, capturing both large anatomical structures and subtle pathological features. The framework incorporates a modified StyleGAN-based generator with dilated convolutions and strategic upsampling adjustments to preserve critical retinal structures while enhancing pathological detail representation. Comprehensive evaluations on the DDR, DRIVE, and IDRiD datasets demonstrate that FundusGAN consistently outperforms state-of-the-art methods across multiple metrics (SSIM: 0.8863, FID: 54.2, KID: 0.0436 on DDR). Furthermore, disease classification experiments reveal that augmenting training data with FundusGAN-generated images significantly improves diagnostic accuracy across multiple CNN architectures (up to 6.49% improvement with ResNet50). These results establish FundusGAN as a valuable foundation model component that effectively addresses data scarcity challenges in ophthalmological AI research, enabling more robust and generalizable diagnostic systems while reducing dependency on large-scale clinical data collection. Copyright © 2025, The Authors. All rights reserved.

关键词： Ophthalmology

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：