检索结果-内蒙古大学图书馆

Multimodal Cross-Attention Mechanism-based Algorithm for Elderly Behavior Monitoring and Recognition

Chinese Journal of Electronics 2025年第1期34卷 309-321页

作者： Hao Liu Zhiquan Feng Qingbei Guo School of Information Science and Engineering Unicersity of Jinan Shandong Provincial Key Laboratory of Network Based Intelligent Computing Unicersity of Jinan

In contrast to the general population, behavior recognition among the elderly poses increased specificity and difficulty, rendering the reliability and usability aspects of safety monitoring systems for the elderly more challenging. Hence, this study proposes a multi-modal perception-based solution for an elderly safety monitoring recognition system. The proposed approach introduces a recognition algorithm based on multi-modal cross-attention mechanism, innovatively incorporating complex information such as scene context and voice to achieve more accurate behavior recognition. By fusing four modalities, namely image, skeleton, sensor data, and audio, we further enhance the accuracy of recognition. Additionally, we introduce a novel human-robot interaction mode, where the system associates directly recognized intentions with robotic actions without explicit commands, delivering a more natural and efficient elderly assistance paradigm. This mode not only elevates the level of safety monitoring for the elderly but also facilitates a more natural and efficient caregiving approach. Experimental results demonstrate significant improvement in recognition accuracy for 11 typical elderly behaviors compared to existing methods.

关键词： Privacy Accuracy Image recognition Speech recognition Robot sensing systems Skeleton Safety Reliability Older adults Monitoring

来源：评论

学校读者我要写书评

暂无评论

MARL-DCF: A Dual-Channel Fusion Routing Algorithm based on Multi-agent Reinforcement Learning 24th

MARL-DCF: A Dual-Channel Fusion Routing Algorithm Based on ...

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Li, Zhenyu Sun, Runyuan Liang, Zhifeng Liu, Bo Sun, Zhenxiang School of Information Science and Engineering University of Jinan Jinan250022 China Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of Jinan Jinan250022 China

ISBN: (纸本)9789819615414

Emerging applications have imposed stringent demands on Internet quality of service (QoS). To ensure that various data streams within different Internet services receive the appropriate QoS, advancements in flow classification technology, software-defined networking (SDN), and programmable network devices have enabled the network to identify user demands and precisely control fine-grained traffic routing quickly. Although significant progress has been made with these technologies, optimizing the transmission paths of business flows to meet diverse QoS requirements in real-time remains a challenge. To address this issue, this paper proposes a dual-channel fusion routing algorithm based on multi-agent reinforcement learning (MARL-DCF). Implemented within an SDN architecture, the algorithm was experimentally validated in a Mininet environment using real network topologies and traffic trajectories. The results demonstrate that MARL-DCF can optimize routing paths according to the network environment, meet various QoS demands, and exhibit strong adaptability and reliability in complex scenarios such as link failures and traffic fluctuations. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

A Lightweight and High-Fidelity Model for Generalized Audio-Driven 3D Talking Face Synthesis 27th

A Lightweight and High-Fidelity Model for Generalized Audi...

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Liu, Shunce Zhong, Yuwei Wang, Huixuan Peng, Jingliang Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of Jinan Jinan China School of Information Science and Engineering University of Jinan Jinan China School of Software and Microelectronics Peking University Beijing China School of Software Shandong University Jinan China

ISBN: (纸本)9783031781247

We present LW-GeneFace, a lightweight and high-fidelity model for generalized audio-driven facial animation, in this paper. We develop this model by reducing the size while maintaining the synthetic quality of GeneFace, an audio-driven facial animation model known for its high fidelity and generalization capabilities. Specifically, we compress the first and the third stages of GeneFace as they dominate the model size. In the first stage, we propose a lightweight version of the WaveNet-based network inspired by MobileNetV3 and DP-block. It utilizes depthwise separable convolution and dual-path feature extraction to compress the network while maintaining effective feature extraction. The shared network structure in the dual-path feature extraction further reduces model complexity and improves training efficiency. In the third stage, we generate realistic 3D renderings at reduced model size by introducing novelties in RAD-NeRF. Technically, we reduce the hash table sizes in the grid-based encoding modules, as well as present a lightweight bottleneck MLP architecture to increase the non-linearity of the model. Experimental results demonstrate that LW-GeneFace achieves state-of-the-art performance with both model size and synthetic quality considered. The source code of LW-GeneFace will be released after acceptance of this paper. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Animation

来源：评论

学校读者我要写书评

暂无评论

MDGCN-Lt:Fair Web API Classification with Sparse and Heterogeneous Data based on Deep GCN

引用

清华大学学报自然科学版（英文版） 2025年第3期30卷 1294-1314页

作者： Boyuan Yan Yankun Zhang Wenwen Gong Haoyang Wan Wenwei Wang Weiyi Zhong Caixia Bu Department of School of Computer Science Qufu Normal UniversityRizhao 276500China Shandong Provincial University Laboratory for Protected Horticulture Weifang University of Science and TechnologyWeifang 261000China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of JinanJinan 250022China Department of Univalsoft Joint Stock Co. Ltd.Weifang 261000China

Developers integrate web Application Programming Interfaces(APIs)into edge applications,enabling data expansion to the edge computing area for comprehensive coverage of devices in that *** develop edge applications,developers search API categories to select APIs that meet specific ***,the accurate classification of APIs becomes critically ***,existing approaches,as evident on platforms like ***,face significant ***,sparsity in API data reduces classification accuracy in works focusing on single-dimensional API ***,the multidimensional and heterogeneous structure of web APIs adds complexity to data mining tasks,requiring sophisticated techniques for effective integration and analysis of diverse data ***,the long-tailed distribution of API data introduces biases,compromising the fairness of classification *** these challenges,we propose MDGCN-Lt,an API classification approach offering flexibility in using multi-dimensional heterogeneous *** tackles data sparsity through deep graph convolutional networks,exploring high-order feature interactions among API ***-Lt employs a loss function with logit adjustment,enhancing efficiency in handling long-tail data *** results affirm our approach's superiority over existing methods.

关键词： bidirectional encoder representations from transformers deep graph convolutional networks logit adjustment web application programming interface classification application programming interface correlation graph

来源：评论

学校读者我要写书评

暂无评论

HybridFL: Hybrid Approach Toward Privacy-Preserving Federated Learning 6th

HybridFL: Hybrid Approach Toward Privacy-Preserving Federate...

引用

6th International Conference on Security and Privacy in New computing Environments, SPNCE 2023

作者： Ali, Sheraz Mamoon, Saqib Usman, Areeba Abidin, Zain ul Zhao, Chuan School of Information Science and Engineering University of Jinan Jinan250022 China School of Computer Science and Engineering Nanjing University of Science and Technology Nanjing210094 China University of Sargodha Sargodha Pakistan Shandong Provincial Key Laboratory of Network-Based Intelligent Computing University of Jinan Jinan250022 China Shandong Provincial Key Laboratory of Software Engineering University of Jinan Jinan250022 China

ISBN: (纸本)9783031736988

In this study, we introduce a novel Hybrid Federated Learning (HybridFL) approach aimed at enhancing privacy and accuracy in collaborative machine learning scenarios. Our methodology integrates Differential Privacy (DP) and secret sharing techniques to address inference risks during training and protect against information leakage in the output model. Drawing inspiration from recent advances, we present a HybridFL framework that combines the strengths of Homomorphic Encryption (HE) and Multi-Party Computation (MPC) to achieve secure computation without the computational overhead of pure HE methods. Our contributions include a privacy-preserving design for Federated Learning (FL) that ensures local data privacy through secret sharing while leveraging DP mechanisms for noise addition. The system offers resilience against unreliable participants and is evaluated using various machine learning models, including Convolutional Neural networks (CNN), Multi-Layer Perceptrons (MLP), and linear regression. Furthermore, we address potential external threats by deploying predictive model outputs as robust services against inference attacks. Experimental results demonstrate improved accuracy and convergence speed, establishing the viability of HybridFL as an effective solution for collaborative machine learning with enhanced privacy guarantees. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Multi-agent reinforcement learning for vibration control of regenerative active suspension

引用

Engineering Applications of Artificial Intelligence 2025年 153卷

作者： Gao, Xiaotian Du, Yu Han, Shiyuan Zhao, Wenxiu Zhou, Jin Zhang, Tong Chen, C.L. Philip School of Artificial Intelligence Shandong Women's University Jinan250399 China Shandong Provincial Key Laboratory of Network-Based Intelligent Computing University of Jinan Jinan250022 China School of Computer Science and Engineering South China University of Technology Guangzhou510641 China

The comfort and responsiveness of active suspension systems have surpassed those of traditional suspension systems, but their development and implementation have been significantly constrained by high power requirements. In response to the current drive for energy conservation and emission reduction, regenerative active suspension systems effectively address the issues of high power consumption and energy loss while demonstrating considerable market potential. The main contribution of the investigation lies in designing a novel regenerative active suspension system specifically for electric vehicles powered by electric motors. The system integrates an electric motor and a generator to ensure precise power delivery and energy recovery from suspension dynamics. To fulfill the dual requirements of energy regeneration and ride smoothness, the Double Proximal Policy Optimization-Multi Dimensional Output (DPPO-MDO) algorithm has been devised based on multi-agent reinforcement learning. The algorithm features two agents equipped with the Proximal Policy Optimization (PPO) algorithm, each managing the operations of the motor and generator respectively. Furthermore, to accommodate diverse scenarios and balance comfort with energy efficiency, economic and comfort modes have been developed for the regenerative active suspension through the DPPO-MDO algorithm. Comprehensive experimental analysis demonstrates that both redesigned reward function and innovative reinforcement learning framework significantly enhance the application of multi-agent reinforcement learning in regenerative active suspension systems. © 2025 Elsevier Ltd

关键词： Active suspension systems

来源：评论

学校读者我要写书评

暂无评论

The Optimal Solution of Multi-kernel Regularization Learning

引用

Acta Mathematica Sinica,English Series 2013年第8期29卷 1607-1616页

作者： Hong Wei SUN Ping LIU School of Mathematical Science University of Ji'nanShandong Provincial Key Laboratory of Network based Intelligent Computing

In regularized kernel methods, the solution of a learning problem is found by minimizing a functional consisting of a empirical risk and a regularization term. In this paper, we study the existence of optimal solution of multi-kernel regularization learning. First, we ameliorate a previous conclusion about this problem given by Micchelli and Pontil, and prove that the optimal solution exists whenever the kernel set is a compact set. Second, we consider this problem for Gaussian kernels with variance σ∈（0,∞）, and give some conditions under which the optimal solution exists.

关键词： Learning theory multi-kernel regularization optimal solution Gaussian kernels

来源：评论

学校读者我要写书评

暂无评论

A versioning-based acceleration method for software online upgrade

引用

Recent Patents on Computer Science 2016年第1期9卷 81-88页

作者： Ma, Kun Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of Jinan Jinan China

Background: Many large software systems composed of lots of files are expected to offer customers 24/7 services. On the other hand, periodic and frequent updates are forced to fix any bugs and generate several revisions. Therefore, how to manage the revisions without human intervention is challenging. The aims of this paper are to introduce readers to one versioning-based acceleration method for software online upgrade, to discuss the details of versioning-based upgrade and download acceleration, and review some recent patents on software online upgrade. Methods: This paper describes Patent Publication Number CN102868731A, titled "Method and appliance for software online upgrade and downloading acceleration", issued by State Intellectual Property Office of the P.R.C. on January 9, 2013. The patent is based on versioning-based acceleration method for software online upgrade. First, client computer system requests distributed hash table (DHT) nodes to obtain update resources. Second, DHT node discovers resource seeds (version control server) or peers (client systems). Finally, resource seeds send the client a list of applicable updates. Results: Versioning-based upgrade and download acceleration are proposed to address the issue of software online upgrade. DHT algorithm is used to accelerate the download process and manage the update contents using version control tools. As a result of P2P network, the update process runs smoothly under single node failure. Although there is room for improvement of this method, the current applicability has been illustrated by efficiently solving software online upgrade problems. Conclusion: The idea and a disclosed embodiment of a patent (Patent CN102868731A, issued by State Intellectual Property Office of the P.R.C.) are presented, which is based on versioning-based acceleration method for software online upgrade. It includes receiving software updates at groups of distributed hash table (DHT) nodes, finding resource seeds (version

关键词： Peer to peer networks

来源：评论

学校读者我要写书评

暂无评论

Automatic literature metadata extraction from datacite services

引用

Recent Patents on Computer Science 2018年第1期11卷 25-31页

作者： Ma, Kun Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of Jinan Jinan250022 China

Background: Generally, it is difficult to obtain the literature metadata in a unified way because the source data of literature is heterogeneous. Researchers developed a series of systems marked by digital object to manage them and obtained a good effect. Though there are several DOI systems, we face with some problems in promoting the use of them. Objective: To address this issue of promoting literature identifier extraction, this paper has proposed automatic literature metadata extraction from DataCite services. Method: This paper describes Patent Publication Number CN103279361A, titled "Method and System for Bookmark-triggered Literature Sharing", issued by State Intellectual Property Office of the P.R.C. on January 27, 2016. A literature metadata extraction system supporting both personal computer and mobile terminal is developed using the integration of DataCite content negotiation, DataCite metadata search, and HTML template extraction. The architecture of this system is divided into model, view, service and controller. An important contribution of this article is to design a cross-platform and universal way to extract digital literature with/without DOI. Results: The analysis of application's effect and piratical test case show the ability to verify the authenticity of automatic literature metadata extraction from DataCite services. The contributions of our method are literature identifier extraction from DOI proxy, template extraction using Roadrunner, and bookmarklet-based literature sharing. Conclusion: The idea and a disclosed embodiment of a patent (Patent CN103279361A, issued by State Intellectual Property Office of the P.R.C.) are presented, which is based on the distribution of literature metadata extraction. In one disclosed embodiment, this method contains literature identifier extraction from DOI proxy, template extraction using Roadrunner, and bookmarklet-based literature sharing. © 2018 Bentham Science Publishers.

关键词： Extraction

来源：评论

学校读者我要写书评

暂无评论

A column-aware data caching method and system

引用

Recent Patents on Computer Science 2018年第3期11卷 161-168页

作者： Ma, Kun Liu, Shuhui Duan, Chunsun Shandong Provincial Key Laboratory of Network Based Intelligent Computing University of Jinan Jinan250022 China

Background: Large-scale data has brought more challenges in the aspects of efficient storage and access requirements. Due merely to differences in the programming interface and database schema, the emerging new database cannot replace RDBMS completely. Therefore, in a longer period in the future, schema-free databases that will assist RDBMS to address the access bottleneck is a broad solution of big data access in industry and academia. Objective: Since schema-free data has the features of high performance and extendibility, it is generally used as the storage of data cache. But there are few effective solutions to keep high cache hit. The frequent access data is not always guaranteed in the cache. Method: This paper describes Patent Publication Number CN103631972A, titled "Method and System for column-aware data caching", issued by the State Intellectual Property Office of the P.R.C. on December 23, 2013. The caching process includes judging cache hit or miss, updating column access frequency, and change data capture. In order to increase the cache hit rate, the patent is related to cache replacement using column access frequency. There are three circumstances to update column access frequency and maintain cache replacement: transactional updates, non-transactional query, and cache listener. Transactional updates will synchronize the updates of the database to the cache system. Non-transactional query and cache listener will rectify the column access frequency using frequency counter. Results: There are four results. Firstly, column-aware data caching has the features of low query time and high throughput. Secondly, dynamic cache replacement using column access frequency improves the cache hit rate and guarantees eventual cache consistency. Thirdly, cache listener can clean the expired data to guarantee the hot data in the cache. Finally, this column-aware data caching system is transparent to the developers. Cache consistency in this paper is slightly different fr

关键词： Big data

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：