检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Pentyala, Sikha Pereira, Mayana De Cock, Martine School of Engineering and Technology University of Washington Tacoma United States Department of Electrical Engineering Uni-versidade de Brasilia Brazil Department of Applied Mathematics Computer Science and Statistics Ghent University Belgium

Data is the lifeblood of the modern world, forming a fundamental part of AI, decision-making, and research advances. With increase in interest in data, governments have taken important steps towards a regulated data world, drastically impacting data sharing and data usability and resulting in massive amounts of data confined within the walls of organizations. While synthetic data generation (SDG) is an appealing solution to break down these walls and enable data sharing, the main drawback of existing solutions is the assumption of a trusted aggregator for generative model training. Given that many data holders may not want to, or be legally allowed to, entrust a central entity with their raw data, we propose a framework for collaborative and private generation of synthetic tabular data from distributed data holders. Our solution is general, applicable to any marginal-based SDG, and provides input privacy by replacing the trusted aggregator with secure multi-party computation (MPC) protocols and output privacy via differential privacy (DP). We demonstrate the applicability and scalability of our approach for the state-of-the-art select-measure-generate SDG algorithms MWEM+PGM and AIM. Copyright © 2024, The Authors. All rights reserved.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

From spacetime effective media to cascaded frequency conversion

From spacetime effective media to cascaded frequency convers...

引用

URSI Atlantic Radio Science Conference (URSI AT-RASC)

作者： Victor Pacheco-Peña Nader Engheta School of Mathematics Statistics and Physics Newcastle University Newcastle Upon Tyne United Kingdom Department of Electrical and Systems Engineering University of Pennsylvania Philadelphia PA USA

ISBN: (数字)9789463968102

ISBN: (纸本)9798350360257

Wave propagation in materials modulated in space and time has become a new paradigm for a full control of wavematter interactions. Here, we will discuss our recent efforts in the field using time interfaces (i.e., a rapid change in time of the electromagnetic parameters of the medium where a wave travels). Applications include the use of time interfaces to enable space-time effective media, as well as frequency conversion in dielectric slab waveguides (single and cascaded waveguides) where the cladding material is time varying while the core does not change with time.

关键词： Waveguide transitions Claddings Propagation Waveguide theory Frequency conversion Nonhomogeneous media Dielectrics

来源：评论

学校读者我要写书评

暂无评论

Dual Adversarial Alignment for Realistic Support-Query Shift Few-shot Learning

arXiv

引用

arXiv 2023年

作者： Jiang, Siyang Fang, Rui Chen, Hsi-Wen Ding, Wei Chen, Ming-Syan School of Mathematics and Statistics Huizhou University Huizhou China Department of Electrical and Engineering National Taiwan University TW Taipei Taiwan

Support-query shift few-shot learning aims to classify unseen examples (query set) to labeled data (support set) based on the learned embedding in a low-dimensional space under a distribution shift between the support set and the query set. However, in real-world scenarios the shifts are usually unknown and varied, making it difficult to estimate in advance. Therefore, in this paper, we propose a novel but more difficult challenge, RSQS, focusing on Realistic Support-Query Shift few-shot learning. The key feature of RSQS is that the individual samples in a meta-task are subjected to multiple distribution shifts in each meta-task. In addition, we propose a unified adversarial feature alignment method called DUal adversarial ALignment framework (DuaL) to relieve RSQS from two aspects, i.e., inter-domain bias and intra-domain variance. On the one hand, for the inter-domain bias, we corrupt the original data in advance and use the synthesized perturbed inputs to train the repairer network by minimizing distance in the feature level. On the other hand, for intra-domain variance, we proposed a generator network to synthesize hard, i.e., less similar, examples from the support set in a self-supervised manner and introduce regularized optimal transportation to derive a smooth optimal transportation plan. Lastly, a benchmark of RSQS is built with several state-of-the-art baselines among three datasets (CIFAR100, mini-ImageNet, and Tiered-Imagenet). Experiment results show that DuaLsignificantly outperforms the state-of-the-art methods in our benchmark. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

EuclidNets: An Alternative Operation for Efficient Inference of Deep Learning Models

引用

SN Computer Science 2023年第5期4卷 1-11页

作者： Li, Xinlin Parazeres, Mariana Oberman, Adam Ghaffari, Alireza Asgharian, Masoud Nia, Vahid Partovi Huawei Noah’s Ark Lab Montreal Research Centre 7101 Park Avenue Montreal H3N 1X9 QC Canada Department of Mathematics and Statistics McGill University 805 Sherbrooke Street West Montreal H3A 0B9 QC Canada Department of Mathematics and Industrial Engineering Polytechnique Montreal 2500 Chem. Polytechnique Montreal H3T 1J4 QC Canada

With the advent of deep learning applications on edge devices, researchers actively try to optimize deep learning model deployment on low-power and restricted memory devices. There are established compression methods such as quantization, pruning, and architecture search that leverage commodity hardware. Apart from conventional compression algorithms, one may redesign the operations of deep learning models, leading to more efficient hardware implementation. To this end, we propose EuclidNet, an efficient computing method designed to be implemented on hardware that replaces multiplication, with squared difference. We show that EuclidNet is aligned with matrix multiplication and can be used as a measure of similarity in the case of convolutional layers. Furthermore, we show that under various transformations and noise scenarios, EuclidNet exhibits the same performance compared to the deep learning models designed with multiplication operations. © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.

关键词： Convolutional neural network Efficient inference Euclidean distance Hardware efficient algorithms

来源：评论

学校读者我要写书评

暂无评论

AutoRank: MCDA Based Rank Personalization for LoRA-Enabled Distributed Learning

arXiv

引用

arXiv 2024年

作者： Chen, Shuaijun Tavallaie, Omid Nazimi, Niousha Chen, Xin Zomaya, Albert Y. School of Computer Science The University of Sydney Australia Department of Engineering Science University of Oxford United Kingdom School of Mathematics and Statistics Central South University China

As data volumes expand rapidly, distributed machine learning has become essential for addressing the growing computational demands of modern AI systems. However, training models in distributed environments is challenging with participants hold skew, Non-Independent-Identically distributed (Non-IID) data. Low-Rank Adaptation (LoRA) offers a promising solution to this problem by personalizing low-rank updates rather than optimizing the entire model, LoRA-enabled distributed learning minimizes computational and maximize personalization for each participant. Enabling more robust and efficient training in distributed learning settings, especially in large-scale, heterogeneous systems. Despite the strengths of current state-of-the-art methods, they often require manual configuration of the initial rank, which is increasingly impractical as the number of participants grows. This manual tuning is not only time-consuming but also prone to suboptimal configurations. To address this limitation, we propose AutoRank, an adaptive rank-setting algorithm inspired by the bias-variance trade-off. AutoRank leverages the MCDA method TOPSIS to dynamically assign local ranks based on the complexity of each participant’s data. By evaluating data distribution and complexity through our proposed data complexity metrics, AutoRank provides fine-grained adjustments to the rank of each participant’s local LoRA model. This adaptive approach effectively mitigates the challenges of double-imbalanced, non-IID data. Experimental results demonstrate that AutoRank significantly reduces computational overhead, enhances model performance, and accelerates convergence in highly heterogeneous federated learning environments. Through its strong adaptability, AutoRank offers a scalable and flexible solution for distributed machine learning. 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

Interpretable Machine Learning Models Can Outperform Statistical Models in Solder Joint Reliability

Interpretable Machine Learning Models Can Outperform Statist...

引用

Intersociety Conference on Thermal and Thermomechanical Phenomena in Electronic Systems (ITHERM)

作者： Qais Qasaimeh Haoran Li Sa’d Hamasha John Evans Jia Liu Department of Industrial and Systems Engineering Auburn University AL Department of Mathematics and Statistics Auburn University AL The Center for Advanced Vehicle and Extreme Environmental Electronics (CAVE3)

ISBN: (数字)9798350364330

ISBN: (纸本)9798350364347

Weibull regression and the Cox proportional hazard model (Cox-PHM) are commonly employed for analyzing the solder joint reliability due to their simplicity and meaningful results. Recently, various machine learning (ML) models have been applied and shown similar or better performance than conventional approaches, yet they do not incur widespread adoption in reliability modeling because of their limited transparency and interpretability. In this research, we compare the predictive capabilities of Weibull regression and Cox-PHM with ML techniques (specifically, Random Survival Forests [RSF] and Gradient Boosting [GB]) using experimental data from an accelerated thermal cycling reliability test and calculating the concordance index (c-index) as an evaluation metric to assess the effectiveness of a model in predicting the sequence of failure times in the tests. The findings with our dataset showed that ML-based models have a similar performance with the traditional Weibull regression (c-index = 0.808), and GB can even outperform with a high c-index value of 0.848. The superior performance of ML models can be attributed to their capacity to capture nonlinearities and intricate relationships within the data. Additionally, we utilized feature importance scores to understand which factors/variables significantly influence the model's prediction. Such interpretable ML methods can provide clear insights into how models make predictions, thereby enhancing trust and the adoption of innovative ML techniques in electronics manufacturing.

关键词： Measurement Forests Thermomechanical processes Life estimation Predictive models Data models Hazards Reliability Indexes Soldering

来源：评论

学校读者我要写书评

暂无评论

DIMON: Learning Solution Operators of Partial Differential Equations on a Diffeomorphic Family of Domains

arXiv

引用

arXiv 2024年

作者： Yin, Minglang Charon, Nicolas Brody, Ryan Lu, Lu Trayanova, Natalia Maggioni, Mauro Department of Biomedical Engineering Johns Hopkins University BaltimoreMD United States Department of Mathematics University of Houston HoustonTX United States Department of Statistics and Data Science Yale University New HavenCT United States Department of Applied Mathematics and Statistics Johns Hopkins University BaltimoreMD United States School of Medicine Johns Hopkins University BaltimoreMD United States Department of Mathematics Johns Hopkins University BaltimoreMD United States

The solution of a PDE over varying initial/boundary conditions on multiple domains is needed in a wide variety of applications, but it is computationally expensive if the solution is computed de novo whenever the initial/boundary conditions of the domain change. We introduce a general operator learning framework, called DIffeomorphic Mapping Operator learNing (DIMON) to learn approximate PDE solutions over a family of domains {Ωθ}θ, that learns the map from initial/boundary conditions and domain Ωθ to the solution of the PDE, or to specified functionals thereof. DIMON is based on transporting a given problem (initial/boundary conditions and domain Ωθ) to a problem on a reference domain Ω0, where training data from multiple problems is used to learn the map to the solution on Ω0, which is then re-mapped to the original domain Ωθ. We consider several problems to demonstrate the performance of the framework in learning both static and time-dependent PDEs on non-rigid geometries;these include solving the Laplace equation, reaction-diffusion equations, and a multiscale PDE that characterizes the electrical propagation on the left ventricle. This work paves the way toward the fast prediction of PDE solutions on a family of domains and the application of neural operators in engineering and precision medicine. Copyright © 2024, The Authors. All rights reserved.

关键词： Linear equations

来源：评论

学校读者我要写书评

暂无评论

A Semi-Adaptive Preservative Scheme for a Fractional Quenching Convective-Diffusion Problem

SSRN

引用

SSRN 2023年

作者： Liu, Nabing Zhu, Lin Sheng, Qin School of Mathematics and Statistics Ningxia University Yinchuan750021 China Department of Mathematics Center for Astrophysics Space Physics and Engineering Research Baylor University WacoTX76798-7328 United States

A preservative scheme is presented and analyzed for the solution of a quenching type convective-diffusion problem modeled through one-sided Riemann-Liouville space-fractional derivatives. Properly weighted Gr\"unwald formulasare employed for the discretization of the fractional derivative. A forward difference approximation is consideredin the approximation of the convective term of the nonlinear equation. Temporal steps are optimized via an asymptotic arc-length monitoring mechanism till the quenching point. Under suitable constraints on spatial-temporal discretization steps, the monotonicity, positivity preservations of the numerical solution and numerical stability of the scheme are *** numerical experiments are designed to demonstrate and simulate key characteristics of the semi-adaptivescheme constructed, including critical length, quenching time and quenching location of the fractional quenching phenomena formulated through the one-sided space-fractional convective-diffusion initial-boundary value *** of the convective function to quenching are also investigated. Computational results obtained are carefully compared with those acquired from conventional integer order quenching convection-diffusion problems for validating anticipated accuracy. The experiments exhibit throughout high validity and feasibility of our preservative method accomplished. © 2023, The Authors. All rights reserved.

关键词： Nonlinear equations

来源：评论

学校读者我要写书评

暂无评论

Multi-objective Frequency Conversion Sinusoidal Chaotic Neural Network

Multi-objective Frequency Conversion Sinusoidal Chaotic Neur...

引用

第43届中国控制会议

作者： Zhiqiang Hu Xiaodong Zhao Xiaoqian Li Lei Wang Zhonghuan Yang Liwei Fei School of Mathematics and Statistics Taishan University School of Automation Beijing Information Science and Technology University Department of Information Engineering Shandong Vocational Institute of Fashion Technology

ISBN: (数字)9789887581581

ISBN: (纸本)9798350366907

For the defect of Frequency Conversion Sinusoidal Chaotic Neural Network(FCSCNN) cannot solve the multi-objective optimization *** analyzing the action mechanism and design principle of the existing multi-objective intelligent optimization algorithm,this paper designs a multi-objective optimization algorithm model(MOFCSCNN) suitable for FCSCNN from three aspects,namely,the selection and update of global optimal solution,extraction of non-dominant solution and management of *** experiments on ZDT series benchmark functions,the feasibility and effectiveness of the algorithm are *** order to further test the performance of the algorithm,it is compared with several commonly used multi-objective optimization *** to the simulation results,MOFCSCNN algorithm can realize the optimization of multi-objective problems,and shows better performance in convergence and optimization speed than several existing algorithms.

关键词： chaotic neural network multi-objective intelligent optimization non-dominant solution archive

来源：评论

学校读者我要写书评

暂无评论

Vaccine efficacy and SARS-CoV-2 control in California and *** the session 2020e2026:A modeling study

引用

Infectious Disease Modelling 2022年第1期7卷 62-81页

作者： Md Shahriar Mahmud Md Kamrujjaman Md Mashih Ibn Yasin Adan Md Alamgir Hossain Md Mizanur Rahman Md Shahidul Islam Muhammad Mohebujjaman Md Mamun Molla Department of Computer Science and Engineering State University of BangladeshDhaka1205Bangladesh Department of Mathematics University of DhakaDhaka1000Bangladesh Department of Mathematics and Statistics University of CalgaryCalgaryABCanada Computational Biology Research Lab(CBRL) Department of PharmacyJagannath UniversityDhaka1100Bangladesh Hitotsubashi Institute for Advanced Study Hitotsubashi UniversityNaka KunitachiTokyo186-8601Japan Department of Mathematics and Physics Texas A&M International UniversityLaredoTX78041USA Department of Mathematics&Physics North South UniversityDhaka1229Bangladesh Center for Applied Scientific Computing(CASC) North South UniversityDhaka1229Bangladesh

Background:Besides maintaining health precautions,vaccination has been the only prevention from SARS-CoV-2,though no clinically proved 100%effective vaccine has been developed till *** this stage,to withhold the debris of this pandemic-experts need to know the impact of the vaccine efficacy rates,the threshold level of vaccine effectiveness and how long this pandemic may extent with vaccines that have different efficacy *** this article,a mathematical model study has been done on the importance of vaccination and vaccine efficiency rate during an ongoing ***:We simulated a five compartment mathematical model to analyze the pandemic scenario in both California,and whole *** considered four vaccines,Pfizer(95%),Moderna(94%),AstraZeneca(79%),and Johnson&Johnson(72%),which are being used rigorously to control the SARS-CoV-2 pandemic,in addition with two special cases:a vaccine with 100%efficacy rate and no vaccine under ***-CoV-2 related data of California,and *** used in this ***:Both the infection and death rates are very high in *** model suggests that the pandemic situation in California will be under control in the last quartile of the year 2023 if vaccination program is continued with the Pfizer *** this time,six waves may happen from the beginning of the immunization where the case fatality and recovery rates will be 1.697%and 98.30%,***,according to the considered model,this period might be extended to the mid of 2024 when vaccines with lower efficacy rates are *** the other hand,the daily cases and deaths in the *** be under control at the end of 2026 with multiple *** the number of susceptible people will fall down to none in the beginning of 2027,there is less chance to stop the vaccination program if vaccinated with a vaccine other than a 100%effective vaccine or Pfizer,and at that case vaccination program must run till the mid of *** to this s

关键词： Vaccine Model California Control measurement SARS-CoV-2

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：