资源类型

期刊论文 322

会议视频 4

年份

2023 65

2022 69

2021 39

2020 39

2019 25

2018 17

2017 22

2016 8

2015 10

2014 1

2013 1

2011 1

2010 1

2009 1

2008 1

2007 3

2006 1

2005 1

2001 3

展开 ︾

关键词

机器学习 27

深度学习 15

人工智能 12

代理模型 2

增材制造 2

强化学习 2

材料设计 2

结构健康监测 2

高层建筑 2

2D—3D配准 1

3D打印 1

6G 1

ARM 1

CAN总线 1

CCUS 1

COVID-19 1

H∞控制；零和动态博弈；强化学习；自适应动态规划；极小极大Q-学习；策略迭代 1

IC产业 1

IgG 1

展开 ︾

检索范围：

排序： 展示方式：

Low-Cost Federated Broad Learning for Privacy-Preserved Knowledge Sharing in the RIS-Aided Internet of

Xiaoming Yuan,Jiahui Chen,Ning Zhang,Qiang Ye,Changle Li,Chunsheng Zhu,Xuemin Sherman Shen,

《工程（英文）》 doi: 10.1016/j.eng.2023.04.015

摘要： High-efficiency and low-cost knowledge sharing can improve the decision-making ability of autonomous vehicles by mining knowledge from the Internet of Vehicles (IoVs). However, it is challenging to ensure high efficiency of local data learning models while preventing privacy leakage in a high mobility environment. In order to protect data privacy and improve data learning efficiency in knowledge sharing, we propose an asynchronous federated broad learning (FBL) framework that integrates broad learning (BL) into federated learning (FL). In FBL, we design a broad fully connected model (BFCM) as a local model for training client data. To enhance the wireless channel quality for knowledge sharing and reduce the communication and computation cost of participating clients, we construct a joint resource allocation and reconfigurable intelligent surface (RIS) configuration optimization framework for FBL. The problem is decoupled into two convex subproblems. Aiming to improve the resource scheduling efficiency in FBL, a double Davidon–Fletcher–Powell (DDFP) algorithm is presented to solve the time slot allocation and RIS configuration problem. Based on the results of resource scheduling, we design a reward-allocation algorithm based on federated incentive learning (FIL) in FBL to compensate clients for their costs. The simulation results show that the proposed FBL framework achieves better performance than the comparison models in terms of efficiency, accuracy, and cost for knowledge sharing in the IoV.

关键词： Knowledge sharing     Internet of Vehicles     Federated learning     Broad learning     Reconfigurable intelligent surfaces     Resource allocation    

HTML PDF 收藏

联邦无监督表示学习 Research Article

张凤达1,况琨1,陈隆1,游兆阳1,沈弢1,肖俊1,张寅1,吴超2,吴飞1,庄越挺1,李晓林3,4,5

《信息与电子工程前沿（英文）》 2023年 第24卷 第8期   页码 1181-1193 doi: 10.1631/FITEE.2200268

摘要： 为利用分布式边缘设备上大量未标记数据，我们在联邦学习中提出一个称为联邦无监督表示学习(FURL)的新问题，以在没有监督的情况下学习通用表示模型，同时保护数据隐私。FURL提出了两个新挑战：（1）客户端之间的数据分布转移（非独立同分布）会使本地模型专注于不同的类别，从而导致表示空间的不一致；（2）如果FURL中客户端之间没有统一的信息，客户端之间的表示就会错位。为了应对这些挑战，我们提出带字典和对齐的联合对比平均(FedCA)算法。FedCA由两个关键模块组成：字典模块，用于聚合来自每个客户端的样本表示并与所有客户端共享，以实现表示空间的一致性；对齐模块，用于将每个客户端的表示与基于公共数据训练的基础模型对齐。我们采用对比方法进行局部模型训练，通过在3个数据集上独立同分布和非独立同分布设定下的大量实验，我们证明FedCA以显著的优势优于所有基线方法。

关键词： 联邦学习；无监督学习；表示学习；对比学习    

HTML PDF 收藏

6G中联邦学习的应用、挑战和机遇 Review

杨照辉,陈明哲,黃繼傑,H. Vincent Poor,崔曙光

《工程（英文）》 2022年 第8卷 第1期   页码 33-41 doi: 10.1016/j.eng.2021.12.002

摘要：

标准的机器学习方法需要在数据中心集中训练数据，从而采用集中式机器学习算法来进行数据分析和推理。然而，由于无线网络中的隐私限制以及无线通信资源受限，边缘设备将数据传输到参数服务器通常是不可取和不切实际的。联邦学习可解决这些问题。联邦学习可以使设备能够在没有数据共享和传输的情况下训练机器学习模型。本文全面概述了未来第六代（6G）无线网络的联邦学习应用。特别是，首先描述了将联邦学习应用于无线通信中的基本要求。然后详细介绍了无线通信中潜在的联邦学习新型应用，讨论了与新型应用相关的主要问题和挑战。最后，描述了用于无线通信的联邦学习的详细实现方案，并给出了联邦学习的难点和应用前景。

关键词： 联邦学习     6G     智能反射面     语义通信     通信感知计算一体化    

HTML PDF 收藏

联邦相互学习：一种针对异构数据、模型和目标的协同机器学习方法 Research Article

沈弢1,张杰2,贾鑫康2,张凤达1,吕喆奇1,况琨1,吴超3,吴飞1

《信息与电子工程前沿（英文）》 2023年 第24卷 第10期   页码 1390-1402 doi: 10.1631/FITEE.2300098

摘要： 联邦学习（FL）是深度学习中的一种新技术，可以让客户端在保留各自隐私数据的情况下协同训练模型。然而，由于每个客户端的数据分布、算力和场景都不同，联邦学习面临客户端异构环境的挑战。现有方法（如FedAvg）无法有效满足每个客户的定制化需求。为解决联邦学习中的异构挑战，本文首先详述了数据、模型和目标（DMO）这3个主要异构来源，然后提出一种新的联邦相互学习（FML）框架。该框架使得每个客户端都能训练一个考虑到数据异构（DH）的个性化模型。在模型异构（MH）问题上，引入一种“模因模型”作为个性化模型与全局模型之间的中介，并且采用深度相互学习（DML）的知识蒸馏技术在两个异构模型之间传递知识。针对目标异构（OH）问题，通过共享部分模型参数，设计针对特定任务的个性化模型，同时，利用模因模型进行相互学习。本研究通过实验评估了FML在应对DMO异构性方面的表现，并与其他常见FL方法在相似场景下进行对比。实验结果表明，FML在处理FL环境中的DMO问题的表现卓越，优于其他方法。

关键词： 联邦学习；知识蒸馏；隐私保护；异构环境    

HTML PDF 收藏

基于联邦边缘学习的梯度量化和带宽分配优化策略 Research Article

刘沛西1,3,江甲沫2,朱光旭3,程磊4,5,蒋伟1,罗武1,杜滢2,王志勤2

《信息与电子工程前沿（英文）》 2022年 第23卷 第8期   页码 1247-1263 doi: 10.1631/FITEE.2100538

摘要： 由于边缘设备有限算力和边缘网络有限的无线资源，利用联邦边缘学习（federated edge learning, FEEL）训练机器学习模型通常非常耗时。

关键词： 联邦边缘学习；量化优化；带宽分配；训练时间最小化    

HTML PDF 收藏

面向物联网的激励感知区块链辅助的智能边缘缓存与计算迁移研究 Article

王倩, 陈思光, 吴蒙

《工程（英文）》 2023年 第31卷 第12期   页码 127-138 doi: 10.1016/j.eng.2022.10.014

摘要：

人工智能的快速发展将物联网推向了一个新阶段，面对数据的爆炸性增长和用户对更高服务质量的迫切需求，边缘计算和缓存被视为富有前景的技术解决手段。然而，边缘节点（Edge Nodes, ENs）中的资源并不是取之不尽的。本文提出了一种面向物联网的激励感知区块链辅助的智能边缘缓存与计算迁移方案，该方案致力于为协作ENs在资源优化和控制方面提供安全和智能的解决方案。具体地，该方案通过联合优化迁移和缓存决策以及计算和通信资源分配，以最大限度地降低EN中完成任务的总成本。此外，为解决上述优化问题，本文设计了区块链激励和贡献联合感知的联邦深度强化学习算法。在本地训练期间，该算法构建了一个激励感知区块链辅助的协作机制，即在安全保障前提下增强ENs参与协作的意愿。同时，提出了一种基于贡献的联邦聚合方法，即基于EN对全局模型性能提升所做贡献来计算其梯度的聚合权重，以提升训练效果。最后，与其它基准方案相比，数值结果证明本文方案具备高效的资源优化效用，同时在降低总成本和缓存性能方面具有显著优势。


 

关键词： 计算迁移     缓存     激励     区块链     联邦深度强化学习    

HTML PDF 收藏

MSWNet: A visual deep machine learning method adopting transfer learning based upon ResNet 50 for municipal

《环境科学与工程前沿（英文）》 2023年 第17卷 第6期 doi: 10.1007/s11783-023-1677-1

摘要：

● MSWNet was proposed to classify municipal solid waste.

关键词： Municipal solid waste sorting     Deep residual network     Transfer learning     Cyclic learning rate     Visualization    

HTML PDF 收藏

Spatial prediction of soil contamination based on machine learning: a review

《环境科学与工程前沿（英文）》 2023年 第17卷 第8期 doi: 10.1007/s11783-023-1693-1

摘要：

● A review of machine learning (ML) for spatial prediction of soil contamination.

关键词： Soil contamination     Machine learning     Prediction     Spatial distribution    

HTML PDF 收藏

Elucidate long-term changes of ozone in Shanghai based on an integrated machine learning method

《环境科学与工程前沿（英文）》 2023年 第17卷 第11期 doi: 10.1007/s11783-023-1738-5

摘要：

● A novel integrated machine learning method to analyze O3 changes is proposed.

关键词： Ozone     Integrated method     Machine learning    

HTML PDF 收藏

Machine learning in building energy management: A critical review and future directions

《工程管理前沿（英文）》 2022年 第9卷 第2期   页码 239-256 doi: 10.1007/s42524-021-0181-1

摘要： Over the past two decades, machine learning (ML) has elicited increasing attention in building energy management (BEM) research. However, the boundary of the ML-BEM research has not been clearly defined, and no thorough review of ML applications in BEM during the whole building life-cycle has been published. This study aims to address this gap by reviewing the ML-BEM papers to ascertain the status of this research area and identify future research directions. An integrated framework of ML-BEM, composed of four layers and a series of driving factors, is proposed. Then, based on the hype cycle model, this paper analyzes the current development status of ML-BEM and tries to predict its future development trend. Finally, five research directions are discussed: (1) the behavioral impact on BEM, (2) the integration management of renewable energy, (3) security concerns of ML-BEM, (4) extension to other building life-cycle phases, and (5) the focus on fault detection and diagnosis. The findings of this study are believed to provide useful references for future research on ML-BEM.

关键词： building energy management     machine learning     integrated framework     knowledge evolution    

HTML PDF 收藏

Using machine learning models to explore the solution space of large nonlinear systems underlying flowsheet

《化学科学与工程前沿（英文）》 2022年 第16卷 第2期   页码 183-197 doi: 10.1007/s11705-021-2073-7

摘要： Flowsheet simulations of chemical processes on an industrial scale require the solution of large systems of nonlinear equations, so that solvability becomes a practical issue. Additional constraints from technical, economic, environmental, and safety considerations may further limit the feasible solution space beyond the convergence requirement. A priori, the design variable domains for which a simulation converges and fulfills the imposed constraints are usually unknown and it can become very time-consuming to distinguish feasible from infeasible design variable choices by simply running the simulation for each choice. To support the exploration of the design variable space for such scenarios, an adaptive sampling technique based on machine learning models has recently been proposed. However, that approach only considers the exploration of the convergent domain and ignores additional constraints. In this paper, we present an improvement which particularly takes the fulfillment of constraints into account. We successfully apply the proposed algorithm to a toy example in up to 20 dimensions and to an industrially relevant flowsheet simulation.

关键词： machine learning     flowsheet simulations     constraints     exploration    

HTML PDF 收藏

Machine learning for fault diagnosis of high-speed train traction systems: A review

《工程管理前沿（英文）》 doi: 10.1007/s42524-023-0256-2

摘要： High-speed trains (HSTs) have the advantages of comfort, efficiency, and convenience and have gradually become the mainstream means of transportation. As the operating scale of HSTs continues to increase, ensuring their safety and reliability has become more imperative. As the core component of HST, the reliability of the traction system has a substantially influence on the train. During the long-term operation of HSTs, the core components of the traction system will inevitably experience different degrees of performance degradation and cause various failures, thus threatening the running safety of the train. Therefore, performing fault monitoring and diagnosis on the traction system of the HST is necessary. In recent years, machine learning has been widely used in various pattern recognition tasks and has demonstrated an excellent performance in traction system fault diagnosis. Machine learning has made considerably advancements in traction system fault diagnosis; however, a comprehensive systematic review is still lacking in this field. This paper primarily aims to review the research and application of machine learning in the field of traction system fault diagnosis and assumes the future development blueprint. First, the structure and function of the HST traction system are briefly introduced. Then, the research and application of machine learning in traction system fault diagnosis are comprehensively and systematically reviewed. Finally, the challenges for accurate fault diagnosis under actual operating conditions are revealed, and the future research trends of machine learning in traction systems are discussed.

关键词： high-speed train     traction systems     machine learning     fault diagnosis    

HTML PDF 收藏

Dynamic prediction of moving trajectory in pipe jacking: GRU-based deep learning framework

《结构与土木工程前沿（英文）》   页码 994-1010 doi: 10.1007/s11709-023-0942-5

摘要： The moving trajectory of the pipe-jacking machine (PJM), which primarily determines the end quality of jacked tunnels, must be controlled strictly during the entire jacking process. Developing prediction models to support drivers in performing rectifications in advance can effectively avoid considerable trajectory deviations from the designed jacking axis. Hence, a gated recurrent unit (GRU)-based deep learning framework is proposed herein to dynamically predict the moving trajectory of the PJM. In this framework, operational data are first extracted from a data acquisition system; subsequently, they are preprocessed and used to establish GRU-based multivariate multistep-ahead direct prediction models. To verify the performance of the proposed framework, a case study of a large pipe-jacking project in Shanghai and comparisons with other conventional models (i.e., long short-term memory (LSTM) network and recurrent neural network (RNN)) are conducted. In addition, the effects of the activation function and input time-step length on the prediction performance of the proposed framework are investigated and discussed. The results show that the proposed framework can dynamically and precisely predict the PJM moving trajectory during the pipe-jacking process, with a minimum mean absolute error and root mean squared error (RMSE) of 0.1904 and 0.5011 mm, respectively. The RMSE of the GRU-based models is lower than those of the LSTM- and RNN-based models by 21.46% and 46.40% at the maximum, respectively. The proposed framework is expected to provide an effective decision support for moving trajectory control and serve as a foundation for the application of deep learning in the automatic control of pipe jacking.

关键词： dynamic prediction     moving trajectory     pipe jacking     GRU     deep learning    

HTML PDF 收藏

Machine learning modeling identifies hypertrophic cardiomyopathy subtypes with genetic signature

《医学前沿（英文）》 2023年 第17卷 第4期   页码 768-780 doi: 10.1007/s11684-023-0982-1

摘要： Previous studies have revealed that patients with hypertrophic cardiomyopathy (HCM) exhibit differences in symptom severity and prognosis, indicating potential HCM subtypes among these patients. Here, 793 patients with HCM were recruited at an average follow-up of 32.78 ± 27.58 months to identify potential HCM subtypes by performing consensus clustering on the basis of their echocardiography features. Furthermore, we proposed a systematic method for illustrating the relationship between the phenotype and genotype of each HCM subtype by using machine learning modeling and interactome network detection techniques based on whole-exome sequencing data. Another independent cohort that consisted of 414 patients with HCM was recruited to replicate the findings. Consequently, two subtypes characterized by different clinical outcomes were identified in HCM. Patients with subtype 2 presented asymmetric septal hypertrophy associated with a stable course, while those with subtype 1 displayed left ventricular systolic dysfunction and aggressive progression. Machine learning modeling based on personal whole-exome data identified 46 genes with mutation burden that could accurately predict subtype propensities. Furthermore, the patients in another cohort predicted as subtype 1 by the 46-gene model presented increased left ventricular end-diastolic diameter and reduced left ventricular ejection fraction. By employing echocardiography and genetic screening for the 46 genes, HCM can be classified into two subtypes with distinct clinical outcomes.

关键词： machine learning methods     hypertrophic cardiomyopathy     genetic risk    

HTML PDF 收藏

A new automatic convolutional neural network based on deep reinforcement learning for fault diagnosis

《机械工程前沿（英文）》 2022年 第17卷 第2期 doi: 10.1007/s11465-022-0673-7

摘要： Convolutional neural network (CNN) has achieved remarkable applications in fault diagnosis. However, the tuning aiming at obtaining the well-trained CNN model is mainly manual search. Tuning requires considerable experiences on the knowledge on CNN training and fault diagnosis, and is always time consuming and labor intensive, making the automatic hyper parameter optimization (HPO) of CNN models essential. To solve this problem, this paper proposes a novel automatic CNN (ACNN) for fault diagnosis, which can automatically tune its three key hyper parameters, namely, learning rate, batch size, and L2-regulation. First, a new deep reinforcement learning (DRL) is developed, and it constructs an agent aiming at controlling these three hyper parameters along with the training of CNN models online. Second, a new structure of DRL is designed by combining deep deterministic policy gradient and long short-term memory, which takes the training loss of CNN models as its input and can output the adjustment on these three hyper parameters. Third, a new training method for ACNN is designed to enhance its stability. Two famous bearing datasets are selected to evaluate the performance of ACNN. It is compared with four commonly used HPO methods, namely, random search, Bayesian optimization, tree Parzen estimator, and sequential model-based algorithm configuration. ACNN is also compared with other published machine learning (ML) and deep learning (DL) methods. The results show that ACNN outperforms these HPO and ML/DL methods, validating its potential in fault diagnosis.

关键词： deep reinforcement learning     hyper parameter optimization     convolutional neural network     fault diagnosis    

HTML PDF 收藏

标题 作者 时间 类型 操作

Low-Cost Federated Broad Learning for Privacy-Preserved Knowledge Sharing in the RIS-Aided Internet of

Xiaoming Yuan,Jiahui Chen,Ning Zhang,Qiang Ye,Changle Li,Chunsheng Zhu,Xuemin Sherman Shen,

期刊论文

联邦无监督表示学习

张凤达1,况琨1,陈隆1,游兆阳1,沈弢1,肖俊1,张寅1,吴超2,吴飞1,庄越挺1,李晓林3,4,5

期刊论文

6G中联邦学习的应用、挑战和机遇

杨照辉,陈明哲,黃繼傑,H. Vincent Poor,崔曙光

期刊论文

联邦相互学习：一种针对异构数据、模型和目标的协同机器学习方法

沈弢1,张杰2,贾鑫康2,张凤达1,吕喆奇1,况琨1,吴超3,吴飞1

期刊论文

基于联邦边缘学习的梯度量化和带宽分配优化策略

刘沛西1,3,江甲沫2,朱光旭3,程磊4,5,蒋伟1,罗武1,杜滢2,王志勤2

期刊论文

面向物联网的激励感知区块链辅助的智能边缘缓存与计算迁移研究

王倩, 陈思光, 吴蒙

期刊论文

MSWNet: A visual deep machine learning method adopting transfer learning based upon ResNet 50 for municipal

期刊论文

Spatial prediction of soil contamination based on machine learning: a review

期刊论文

Elucidate long-term changes of ozone in Shanghai based on an integrated machine learning method

期刊论文

Machine learning in building energy management: A critical review and future directions

期刊论文

Using machine learning models to explore the solution space of large nonlinear systems underlying flowsheet

期刊论文

Machine learning for fault diagnosis of high-speed train traction systems: A review

期刊论文

Dynamic prediction of moving trajectory in pipe jacking: GRU-based deep learning framework

期刊论文

Machine learning modeling identifies hypertrophic cardiomyopathy subtypes with genetic signature

期刊论文

A new automatic convolutional neural network based on deep reinforcement learning for fault diagnosis

期刊论文