检索结果-内蒙古大学图书馆

2024 International Conference on Optics, Electronics, and Communication engineering, OECE 2024

作者： Li, Zhixin Long, Yang Miao, Jiahao School of Computer Technology and Engineering Changchun Institute of Technology Jilin Changchun130012 China School of Information and Control Engineering Jilin Institute of Chemical Technology Jilin Jilin132022 China

ISBN: (纸本)9781510685437

Recently, the vision transformer (ViT) model of deep learning has achieved surprising performance in the field of computer vision and has been widely used in IoT edge devices. However, the training of ViT models requires a large amount of data and computing resources, which is a challenge for resource-constrained edge IoT devices. To solve the above problems, this paper proposed a lightweight ViT method based on transfer learning. The primary concept of this method is to train large-scale ViT models in the cloud (CloudViT) and deploying small-scale ViT models at the edge (EdgeViT). Firstly, through the method of transfer learning, some underlying parameters of CloudViT were utilized to construct EdgeViT. The purpose is to enable EdgeViT to learn from CloudViT, acquiring knowledge and improving its performance. Secondly, adding a randomly initialized LayerNorm layer before the MLPHead during the training process of EdgeViT, it can improve further model performance. Finally, Experiment results demonstrated that EdgeViT could achieve 91.3% of CloudViT’s performance with only half the parameters and floating-point operations (FLOPs). Moreover, finetuning EdgeViT with a 60% reduction in training time still allows it to achieve 81.3% of CloudViT’s performance. Relevant conclusions can provide technical support for the proposed method. © 2024 SPIE.

关键词： Data modeling Clouds Internet of things Performance modeling Instrument modeling Neural networks Head Visual process modeling Image classification

来源：评论

学校读者我要写书评

暂无评论

PIM-ST: a New Paraphrase Identification Model Incorporating Sequence and Topic Information 4

PIM-ST: a New Paraphrase Identification Model Incorporating ...

引用

4th International Symposium on computer Technology and Information Science, ISCTIS 2024

作者： Mao, Kaili Zhao, Qingjuan Beijing Institute of Control Engineering Beijing China University Beijing Computer School Beijing Information Science and Technology Beijing China

ISBN: (纸本)9798350354560

Paraphrase identification, with the objective to determine whether two sentences are paraphrases of each other, has fostered many applications including natural language inference and document retrieval. Traditional machine learning models for paraphrase identification are generally based on hand-crafted features. Recent deep neural network models generally take semantic vectors as inputs and generate classification results directly. However, existing solutions still have much room for improvement. First, existing models do not fully exploit sequence information (i.e., the order of words) and topic information, which are important for text under-standing. Second, although deep models achieve promising results, they require a large amount of labelled data and have low interpretability. To cope with these problems, we propose a new model called PIM-ST (Paraphrase Identification Model Incorporating Sequence and Topic Information). Specifically, besides semantic information, PIM-ST extracts sequence information and topic information, and adopts a simple machine learning model LightGBM as the classifier. We conduct comparative experiments and ablation studies on two public datasets Microsoft Research Paraphrase Corpus (MSRP) and Quora Question Pairs. Experimental results show that PIM-ST achieves better results than existing state-of-the-art models. © 2024 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Improved SqueezeNet-Based Method for Automotive Paint Film Defect Detection

Improved SqueezeNet-Based Method for Automotive Paint Film D...

引用

2024 International Conference on Artificial Intelligence and Digital Technology, ICAIDT 2024

作者： Yan, Shengnan Shen, Yihui Minjiang University School of Computer and Control Engineering Fujian China Fujian Medical University College of Arts and Sciences Fujian China

ISBN: (纸本)9798350386905

To address the issues of lacking datasets and low recognition accuracy for paint film defects, this paper proposes a denoising diffusion implicit model (DDIM) for data augmentation of paint film defects and innovatively suggests a classification method combining Selective Kernel Networks (SKNet) with SqueezeNet model. Initially, the DDIM is used for dataset expansion, followed by an evaluation of the similarity between generated and captured paint film defect images using multiscale structural similarity (MS-SSIM). The training and generation effects of DDIM are then compared with those of DCGAN. In the SqueezeNet model, a Selective Kernel module is added following the fire7 module to enhance the model's attention mechanism. The results show that all types of paint film defect images generated by DDIM have MS-SSIM indices above 0.64, with most exceeding 0.7. The combined approach of Selective Kernel Networks(SKNet) and SqueezeNet outperforms other attention mechanisms, achieving an accuracy above 96.3%. The method demonstrates promising prospects for paint film defect detection, enhancing identification efficiency and accuracy while reducing detection costs, and is applicable to mobile or embedded devices. © 2024 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Changes in physiology, antioxidant system, and gene expression in Microcystis aeruginosa under fenoxaprop-p-ethyl stress

引用

Environmental Science and Pollution Research 2024年第19期31卷 28754-28763页

作者： Liu, Sijia Ni, Jiawei Guan, Ying Tao, Jianwei Wu, Liang Hou, Meifang Wu, Shichao Xu, Wenwu Zhang, Chu Ye, Jing School of Chemical and Environmental Engineering Shanghai Institute of Technology Shanghai201418 China Los Angeles Regional Water Quality Control Board Los AngelesCA90013 United States School of Ecological Technology and Engineering Shanghai Institute of Technology Shanghai201418 China School of Computer Science and Information Engineering Shanghai Institute of Technology Shanghai201418 China

Fenoxaprop-p-ethyl (FE) is one of the typical aryloxyphenoxypropionate herbicides. FE has been widely applied in agriculture in recent years. Human health and aquatic ecosystems are threatened by the cyanobacteria blooms caused by Microcystis aeruginosa, which is one of the most common cyanobacteria responsible for freshwater blooming. Few studies have been reported on the physiological effects of FE on M. aeruginosa. This study analyzed the growth curves, the contents of chlorophyll a and protein, the oxidative stress, and the microcystin-LR (MC-LR) levels of M. aeruginosa exposed to various FE concentrations (i.e., 0, 0.5, 1, 2, and 5 mg/L). FE was observed to stimulate the cell density, chlorophyll a content, and protein content of M. aeruginosa at 0.5- and 1-mg/L FE concentrations but inhibit them at 2 and 5 mg/L FE concentrations. The superoxide dismutase and catalase activities were enhanced and the malondialdehyde concentration was increased by FE. The intracellular (intra-) and extracellular (extra-) MC-LR contents were also affected by FE. The expression levels of photosynthesis-related genes psbD1, psaB, and rbcL varied in response to FE exposure. Moreover, the expressions of microcystin synthase-related genes mcyA and mcyD and microcystin transportation-related gene mcyH were significantly inhibited by the treatment with 2 and 5 mg/L FE concentrations. These results might be helpful in evaluating the ecotoxicity of FE and guiding the rational application of herbicides in modern agriculture. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Gene expression

来源：评论

学校读者我要写书评

暂无评论

RFNet: A Redundancy-Reduced and Frequency-Aware Feature Fusion Network for RGB-T Semantic Segmentation 10

RFNet: A Redundancy-Reduced and Frequency-Aware Feature Fusi...

引用

10th IEEE Smart World Congress, SWC 2024

作者： Guo, Yixin Chen, Zhenxue Rong, Xuewen Liu, Chengyun Li, Yidi Shandong University School of Control Science and Engineering Jinan China Taiyuan University of Technology Computer Science and Technology Taiyuan China

ISBN: (纸本)9798331520861

Traditional semantic segmentation approaches primarily utilize RGB images, which struggle in complex scenes. To address this challenge, the multi-modal solution that fuses RGB and thermal (RGB-T) images can exploit the advantages of these modalities to improve accuracy and stability of segmentation. However, how to effectively fuse multi-modal and multi-level features remains a significant challenge. To this end, a Redundancy-Reduced and Frequency-Aware Feature Fusion Network (RFNet) is proposed. The RFNet employs a dual-stream encoder-decoder architecture to separately handle the extracted features from two modalities. This network reduces redundancy in shallow feature fusion and integrates frequency-aware attention mechanisms in deep feature fusion, efficiently capturing complementary information from multi-modal and multi-level features. Experiments on the MFNet dataset indicate that RFNet outperforms the majority of current networks in terms of segmentation performance. © 2024 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

STFormer: A dual-stage transformer model utilizing spatio-temporal graph embedding for multivariate time series forecasting

引用

Journal of Intelligent and Fuzzy Systems 2024年第3期46卷 6951-6967页

作者： Xiao, Yuteng Liu, Zhaoyang Yin, Hongsheng Wang, Xingang Zhang, Yudong Jinan China Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China Xuzhou University of Technology Xuzhou China School of Information and Control Engineering China University of Mining and Technology Xuzhou China School of Computing and Mathematical Sciences University of Leicester LeicesterLE1 7RH United Kingdom School of Computer Science and Engineering Southeast University Jiangsu Nanjing China

Multivariate Time Series (MTS) forecasting has gained significant importance in diverse domains. Although Recurrent Neural Network (RNN)-based approaches have made notable advancements in MTS forecasting, they do not effectively tackle the challenges posed by noise and unordered data. Drawing inspiration from advancing the Transformer model, we introduce a transformer-based method called STFormer to address this predicament. The STFormer utilizes a two-stage Transformer to capture spatio-temporal relationships and tackle the issue of noise. Furthermore, the MTS incorporates adaptive spatio-temporal graph structures to tackle the issue of unordered data specifically. The Transformer incorporates graph embedding to combine spatial position information with long-term temporal connections. Experimental results based on typical finance and environment datasets demonstrate that STFormer surpasses alternative baseline forecasting models and achieves state-of-the-art results for single-step horizon and multistep horizon forecasting. © 2024 - IOS Press. All rights reserved.

关键词： Recurrent neural networks

来源：评论

学校读者我要写书评

暂无评论

A Concept Drift Tolerant Abnormal Detection Method for Network Traffic 15

A Concept Drift Tolerant Abnormal Detection Method for Netwo...

引用

15th International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery, CyberC 2023

作者： Wang, Xu Han, Li Sun, Shimin Xu, Guowei Tianjin University of Technology National Engineering Laboratory for Computer Virus Prevention and Control Technology School of Computer Science and Engineering Tianjin China Tiangong University School of Software Tianjin China

ISBN: (纸本)9798350308693

we propose a cutting-edge solution that leverages passive adaptive methods based on ensemble learning to effectively detect anomalous traffic in data streams. Our approach tackles the issue of concept drift by integrating various deep learning models, each offering a unique perspective to learn data features. By doing so, our solution can accurately capture the distribution of data concepts and propose weight layers to make the model adaptable to changes caused by concept drift. In a series of experiments conducted on public datasets, our model demonstrated remarkable effectiveness in detecting anomalies, achieving accuracy rates of over 97% after adapting to concept drift. © 2023 IEEE.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

GRU combined model based on multi-objective optimization for short-term residential load forecasting

引用

Journal of Intelligent and Fuzzy Systems 2024年第4期46卷 10423-10440页

作者： Yi, Lingzhi Peng, Xinlong Fan, Chaodong Wang, Yahui Li, Yunfan Liu, Jiangyong College of Automation and Electronic Engineering Xiangtan University Hunan Engineering Research Center of Multi-energy Cooperative Control Technology Hunan Xiangtan China School of Computer Science and Technology Hainan University Hainan China College of Electrical and Information Engineering Hunan University Changsha China

Reliable and accurate short-term forecasting of residential load plays an important role in DSM. However, the high uncertainty inherent in single-user loads makes them difficult to forecast accurately. Various traditional methods have been used to address the problem of residential load forecasting. A single load forecast model in the traditional method does not allow for comprehensive learning of data characteristics for residential loads, and utilizing RNNs faces the problem of long-term memory with vanishing or exploding gradients in backpropagation. Therefore, a gated GRU combined model based on multi-objective optimization is proposed to improve the short-term residential load forecasting accuracy in this paper. In order to demonstrate the effectiveness, GRUCC-MOP is first experimentally tested with the unimproved model to verify the model performance and forecasting effectiveness. Secondly the method is evaluated experimentally with other excellent forecasting methods: models such as DBN, LSTM, GRU, EMD-DBN and EMD-MODBN. By comparing simulation experiments, the proposed GRU combined model can get better results in terms of MAPE on January, April, July, and November load data, so this proposed method has better performance than other research methods in short-term residential load forecasting. © 2024 - IOS Press. All rights reserved.

关键词： Multiobjective optimization

来源：评论

学校读者我要写书评

暂无评论

A dynamic bidirectional heuristic trust path search algorithm

引用

CAAI Transactions on Intelligence Technology 2022年第3期7卷 340-353页

作者： Jiaying Che Xiangrong Tong Lei Yu School of Computer and Control Engineering Yantai UniversityYantaiShandongChina Department of Computer Science State University of New York at BinghamtonBinghamtonNew YorkUSA

Online social networks greatly promote peoples'online interaction,where trust plays a crucial *** prediction with trust path search is widely used to help users find the trusted friends and obtain valid ***,the shortcomings of accuracy and time still exist in some famous ***,the dynamic bidirectional heuristic search(DBHS)algorithm is proposed in this paper to find the reliable trust path by studying the heuristic ***,the trust value and path length are comprehensively considered to find the most trusted ***,it constrains the traversal depth based on the‘small world’theory and obtains the acceptable path set by using the relaxation coefficientλto relax the depth of the shortest *** this way,some longer path with the higher trust can be considered to improve the precision of ***,an adjustment factor is designed based on the meet in the middle(MM)algorithm to assign search weights to two directions based on the size of the search tree expanded,so as to improve the problem of no priori when fixed parameters are ***,the complexity of unidirectional trust path search can also be reduced by searching from two directions,which can reduce the depth and improve the efficiency of ***,the predictive trust degree is outputted by the trust propagation *** public datasets are used to generate experimental results,which show that DBHS can quickly search and form reliable trust relationship,and it partly improves other algorithms.

关键词： directions algorithm dynamic

来源：评论

学校读者我要写书评

暂无评论

Semi-asynchronous personalized federated learning for short-term photovoltaic power forecasting

引用

Digital Communications and Networks 2023年第5期9卷 1221-1229页

作者： Weishan Zhang Xiao Chen Ke He Leiming Chen Liang Xu Xiao Wang Su Yang College of Computer Science and Technology China University of Petroleum(East China)China Sichuan Energy Internet Research Institute Tsinghua UniversityChina School of Computer and Communication Engineering Beijing University of Science and TechnologyChina State Key Laboratory of Management and Control for Complex Systems Institute of AutomationChinese Academy of SciencesBeijingChina School of Computer Science Fudan UniversityChina

Accurate forecasting for photovoltaic power generation is one of the key enablers for the integration of solar photovoltaic systems into power *** deep-learning-based methods can perform well if there are sufficient training data and enough computational ***,there are challenges in building models through centralized shared data due to data privacy concerns and industry *** learning is a new distributed machine learning approach which enables training models across edge devices while data reside *** this paper,we propose an efficient semi-asynchronous federated learning framework for short-term solar power forecasting and evaluate the framework performance using a CNN-LSTM *** design a personalization technique and a semi-asynchronous aggregation strategy to improve the efficiency of the proposed federated forecasting *** evaluations using a real-world dataset demonstrate that the federated models can achieve significantly higher forecasting performance than fully local models while protecting data privacy,and the proposed semi-asynchronous aggregation and the personalization technique can make the forecasting framework more robust in real-world scenarios.

关键词： Photovoltaic power forecasting Federated learning Edge computing CNN-LSTM

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：