检索结果-内蒙古大学图书馆

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Zhang, WenZheng Hu, Yang Shi, Jing Bai, Xiaoying School of Computer Science Peking University China Center for Information Research Academy of Military Sciences China Advanced Institute of Big Data China

ISBN: (纸本)157735897X

Scaling Deep Neural Networks (DNNs) requires significant computational resources in terms of GPU quantity and compute capacity. In practice, there usually exists a large number of heterogeneous GPU devices due to the rapid release cycle of GPU products. It is highly needed to efficiently and economically harness the power of heterogeneous GPUs, so that it can meet the requirements of DNN research and development. The paper introduces Poplar, a distributed training system that extends Zero Redundancy Optimizer (ZeRO) with heterogeneous-aware capabilities. We explore a broader spectrum of GPU heterogeneity, including compute capability, memory capacity, quantity and a combination of them. In order to achieve high computational efficiency across all heterogeneous conditions, Poplar conducts fine-grained measurements of GPUs in each ZeRO stage. We propose a novel batch allocation method and a search algorithm to optimize the utilization of heterogeneous GPUs clusters. Furthermore, Poplar implements fully automated parallelism, eliminating the need for deploying heterogeneous hardware and finding suitable batch size. Extensive experiments on three heterogeneous clusters, comprising six different types of GPUs, demonstrate that Poplar achieves a training throughput improvement of 1.02 ∼ 3.92x over current state-of-the-art heterogeneous training systems. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Deep neural networks

来源：评论

学校读者我要写书评

暂无评论

Unified Image Harmonization with Region Augmented Attention Normalization

引用

Annals of data science 2024年第5期11卷 1865-1886页

作者： Hou, Junjie Zhang, Yuqi Su, Duo Sino-Danish College University of Chinese Academy of Sciences Zhongguancun East Rd Beijing100190 China School of Mathematical Sciences University of Chinese Academy of Sciences Zhongguancun East Rd Beijing100190 China School of Computer Science and Technology University of Chinese Academy of Sciences Zhongguancun East Rd Beijing100190 China Research Center on Fictitious Economy and Data Science Chinese Academy of Sciences Zhongguancun East Rd Beijing100190 China Key Laboratory of Big Data Mining and Knowledge Management Chinese Academy of Sciences Zhongguancun East Rd Beijing100190 China

The image harmonization task endeavors to adjust foreground information within an image synthesis process to achieve visual consistency by leveraging background information. In academic research, this task conventionally involves the utilization of simple synthesized images and matching masks as inputs. However, obtaining precise masks for image harmonization in practical applications poses a significant challenge, thereby creating a notable disparity between research findings and real-world applicability. To mitigate this disparity, we propose a redefinition of the image harmonization task as "Unified Image Harmonization," where the input comprises only a single image, thereby enhancing its applicability in real-world scenarios. To address this challenge, we have developed a novel framework. Within this framework, we initially employ inharmonious region localization to detect the mask, which is subsequently utilized for harmonization tasks. The pivotal aspect of the harmonization process lies in normalization, which is accountable for information transfer. Nonetheless, the current background-to-foreground information transfer and guidance mechanisms are limited by single-layer guidance, thereby constraining their effectiveness. To overcome this limitation, we introduce Region Augmented Attention Normalization (RA2N), which enhances the attention mechanism for foreground feature alignment, consequently leading to improved alignment and transfer capabilities. Through qualitative and quantitative comparisons on the iHarmony4 dataset, our model exhibits exceptional performance not only in unified image harmonization but also in conventional image harmonization tasks. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

4D foetal cardiac ultrasound image detection based on deep learning with weakly supervised localisation for rapid diagnosis of evolving hypoplastic left heart syndrome

引用

CAAI Transactions on Intelligence Technology 2024年第6期9卷 1485-1499页

作者： Gang Wang Weisheng Li Mingliang Zhou Haobo Zhu Guang Yang Choon Hwai Yap Chongqing Key Laborotory of Image Rocognition Chongqing University of Posts and TelecommunicationsChongqingChina School of Computing and Data Engineering NingboTech UniversityNingboChina Department of Bioengineering Imperial College LondonLondonUK Key Laboratory of Cyberspace Big Data Intelligent Security(Chongqing University of Posts and Telecommunications) Ministry of EducationChongqingChina School of Computer Science Chongqing UniversityChongqingChina University of Oxford OxfordUK Cardiovascular Research Centre Royal Brompton HospitalLondonUK National Heart and Lung Institute Imperial College LondonLondonUK

Hypoplastic left heart syndrome(HLHS)is a rare,complex,and incredibly foetal congenital heart *** decrease neonatal mortality,evolving HLHS(eHLHS)in pregnant women should be critically diagnosed as soon as ***,diagnosis is currently heavily dependent on skilled medical professionals using foetal cardiac ultrasound images,making it difficult to rapidly and easily examine for this ***,the authors propose a cost-effective deep learning framework for rapid diagnosis of eHLHS(RDeH),which we have named ***,the framework implements a coarseto-fine two-stage detection approach,with a structure classification network for 4D human foetal cardiac ultrasound images from various spatial and temporal domains,and a fine detection module with weakly-supervised localisation for high-precision nidus localisation and physician *** experiments extensively compare the authors’network with other state-of-the-art methods on a 4D human foetal cardiac ultrasound image dataset and show two main benefits:(1)it achieved superior average accuracy of 99.37%on three categories of foetal ultrasound images from different cases;(2)it demonstrates visually fine detection performance with weakly supervised *** framework could be used to accelerate the diagnosis of eHLHS,and hence significantly lessen reliance on experienced medical physicians.

关键词： 4D deep learning fetal cardiac ultrasound hypoplastic left heart syndrome weakly-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Intelligent Smishing Detection System for Massive Users 9

Intelligent Smishing Detection System for Massive Users

引用

9th IEEE International Conference on Cloud Computing and Intelligence Systems, CCIS 2023

作者： Li, Qiang Zhang, Jianxin Zhou, Nan Ming, Liang Zhao, Bo Du, Yuejin Zhong, Li Li, Jia Beijing Qihoo Technology Co. Ltd Beijing100015 China National Engineering Research Center of Big Data Collaborative Security Technology Beijing100015 China WuHan University School of Cyber Science and Engineering WuHan430064 China

ISBN: (纸本)9798350304428

Smishing has always been an important means of black-gray drainage. Currently, interception measures are primarily deployed at the network communication layer of SMS transmission. In response, black-gray industry organizations continuously improve their evasion techniques to bypass the interception strategies along the communication chain. This article presents a novel phishing text message interception method that involves collaboration between the terminal and the cloud. Its advantage lies in the combination of deep learning models for text classification and the identification of reserved information. Firstly, the optimization process involves leveraging a massive corpus in the vertical domain, consisting of hundreds of millions of samples, to optimize the Bert model and TextCNN network. This optimization includes incorporating proprietary features such as statistics, pinyin, and glyph. As a result, classification models are specifically designed for both the cloud and terminals, enabling efficient recognition of ordinary phishing text messages. Secondly, to address spear-phishing attacks and various types of spoofing in phishing text messages, an enhanced named entity recognition technology is employed. This technology allows for the extraction of URLs, QQ numbers, WeChat IDs, mobile phone numbers, and other suspected fraudulent reserved information from the text messages. Subsequently, dedicated models for feature fusion of URLs and heuristic discrimination strategies for network accounts are established based on this extracted information. Combining the two aforementioned detection methods, the paper has implemented an intelligent phishing text message detection system, which has been deployed and applied in a real environment with of hundred millions users. The system achieved a recall rate of 98.32% and an accuracy rate of 99.44%. © 2023 IEEE.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Causal fMRI-Mamba: Causal State Space Model for Neural Decoding and Brain Task States Recognition

Causal fMRI-Mamba: Causal State Space Model for Neural Decod...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Weihao Deng Fei Han Qinghua Ling Qing Liu Henry Han School of Computer Science and Communication Engineering JiangSu University China Jiangsu Engineering Research Center of Big Data Ubiquitous Perception and Intelligent Agriculture Applications China School of Computer Jiangsu University of Science and Technology China School of Electronic and Information Engineering West Anhui University China School of Engineering and Computer Science Baylor University USA

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Deep learning advances neural decoding in functional magnetic resonance imaging (fMRI) tasks with convolution and attention-based methods. However, these methods struggle with capturing global spatiotemporal information due to high dimensionality, noise and inter-individual difference of fMRI, which also increase computational complexity and prior bias. To this end, a novel causal state space model, Causal fMRI-Mamba, is proposed for neural decoding and task state mapping. It effectively captures global spatiotemporal information via eliminating local redundancies and capturing long-distance dependencies. Meanwhile, a causal representation framework is designed to extract invariant high-order features and disentangle related causal features, enhancing model performance. Furthermore, a dense connection module is expanded to prevent significant causal information loss in hidden states of inter layers. On the HCP brain task state classification task, Causal fMRI-Mamba achieves better performance and generalization than comparison methods.

关键词： Deep learning Perturbation methods Redundancy Noise Functional magnetic resonance imaging Brain modeling Feature extraction Spatiotemporal phenomena Decoding Speech processing

来源：评论

学校读者我要写书评

暂无评论

A Hybrid Model Combined with SVM and CNN for Community Content Classification 23

A Hybrid Model Combined with SVM and CNN for Community Conte...

引用

23rd IEEE International Conference on High Performance Computing and Communications, 7th IEEE International Conference on data science and Systems, 19th IEEE International Conference on Smart City and 7th IEEE International Conference on Dependability in Sensor, Cloud and big data Systems and Applications, HPCC-DSS-SmartCity-DependSys 2021

作者： Ye, Yukui Xie, Xia Jin, Hai Wang, Duoqiang National Engineering Research Center for Big Data Technology and System School of Computer Science and Technology Huazhong University of Science and Technology Services Computing Technology and System Lab Cluster and Grid Computing Lab Wuhan China School of Computer Science and Technology Hainan University Haikou China

ISBN: (纸本)9781665494571

Community websites bring many conveniences to people, and the classification of community content is playing an important role in website management and information searching. As the carrier of community content, posts are difficult to classify manually. According to the characteristics of community content, a hybrid classification model for machine learning is proposed. This model consists of three steps. Firstly, aiming at the problem of fewer features of posts, a weighted word vector is proposed to enrich the features of posts. Secondly, since the single kernel function of SVM can not completely match all data distributions, a mixed kernel function is employed to improve the model. Finally, in order to fully utilize the powerful feature extraction ability of Convolutional Neural Network as well as the classification ability of SVM, a hybrid model is designed and implemented by replacing softmax layer with SVM classifier. The corresponding experiment results indicate that compared with traditional Convolutional Neural Network, the proposed hybrid model has better performance and stability with classification accuracy improved from 0.9% to 1.4% in general. © 2021 IEEE.

关键词： Support vector machines Smart cities Support vector machine classification Machine learning Feature extraction Stability analysis data models

来源：评论

学校读者我要写书评

暂无评论

Cooperative Relationship Prediction between Scholars in Heterogeneous Academic Network 23

Cooperative Relationship Prediction between Scholars in Hete...

引用

作者： Shi, Jia Jin, Hai Xie, Xia National Engineering Research Center for Big Data Technology and System School of Computer Science and Technology Huazhong University of Science and Technology Services Computing Technology and System Lab Cluster and Grid Computing Lab Wuhan China School of Computer Science and Technology Hainan University Haikou China

ISBN: (纸本)9781665494571

The real academic network belongs to a heterogeneous network, therefore, for the link prediction tasks, some information on the network may be lost if only using homogeneous network methods. In order to make good use of the rich semantic information in a heterogeneous network for better predicting the cooperative relationship between scholars, this paper proposes a cooperative relationship prediction (CoRP) model. The CoRP model includes the following three steps. Firstly, a structural feature extraction module is designed based on meta path by using the similarity metric function to measure the meta path similarity between scholars. Secondly, to overcome the defects that the existing methods usually ignore the attribute information on nodes, three attribute feature presentation functions are constructed to fully utilize the attribute information of scholars in the academic network. In the end, a machine learning classifier which combines the structural similarity and attribute similarity is utilized to learn each attribute feature and complete the task of cooperative relationship prediction. The experimental results reveal that compared with the benchmark method, the proposed method has better performance and interpretability with improvement from 3% to 10% of AUC value for the AMiner academic data set. © 2021 IEEE.

关键词： Measurement Smart cities Semantics Machine learning Benchmark testing Predictive models Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Seastar: A Cache-Efficient and Load-Balanced Key-Value Store on Disaggregated Memory

Seastar: A Cache-Efficient and Load-Balanced Key-Value Store...

引用

2024 IEEE International Conference on Cluster Computing, CLUSTER 2024

作者： Du, Jingwen Wang, Fang Feng, Dan Zeng, Dexin Yi, Sheng Engineering Research Center of Data Storage Systems and Technology Ministry of Education of China Wuhan National Laboratory for Optoelectronics Key Laboratory of Information Storage System China School of Computer Science and Technology Huazhong University of Science and Technology China Shenzhen Huazhong University of Science and Technology Research Institute China

ISBN: (纸本)9798350358711

In modern datacenters, memory disaggregation un-packs monolithic servers to build independent network-connected compute and memory pools, greatly improving resource uti-lization. Existing memory-disaggregated key-value stores adopt ownership sharing or ownership partitioning for request scheduling and they all show poor performance and scalability. The former generates multiple cache misses and multiple network round trips due to poor cache locality, while the latter fails to fully use resources on all Compute Nodes (CNs) due to load imbalance. In this paper, we propose Seastar, a fast and scalable key-value store that achieves high cache locality as well as load balancing. To enable high concurrency and ensure linearizability, Seastar exploits a self-verification version chain to provide out-of-place atomic updates and ensure consistent reads. To improve cache locality, Seastar leverages a pair-based ownership assigning scheme to assign the processing permission of requests corre-sponding to each key to a CN pair through two independent hash functions. To ensure load balancing among CNs, Seastar proposes a decentralized power-of-two-choices routing scheme to route the requests to the less-loaded CNs in CN pairs for handling. Experimental results demonstrate that Seastar improves the throughput by up to 6.7x and significantly reduces the latency compared with state-of-the-art memory-disaggregated key-value stores. © 2024 IEEE.

关键词： Scalability

来源：评论

学校读者我要写书评

暂无评论

Automated Extraction of Active Offshore Oil/Gas Platforms Based on High-Spatial Resolution SDGSAT-1/GLI data: A Case Study in the Northern Gulf of Mexico 2

Automated Extraction of Active Offshore Oil/Gas Platforms Ba...

引用

2023 2nd International Conference on Environmental Remote Sensing and Geographic Information Technology, ERSGIT 2023

作者： Chen, Zhongzhong Yan, Lin Qiu, Yujia Liang, Yueji Li, Xiao-Ming College of Geomatics and Geoinformation Guilin University of Technology Guilin541004 China International Research Center of Big Data for Sustainable Development Goals Beijing100094 China Key Laboratory of Digital Earth Science Aerospace Information Research Institute Chinese Academy of Sciences Beijing100094 China School of Computer and Information Engineering Henan University Kaifeng475001 China

ISBN: (纸本)9781510672949

The acquisition of spatial distribution information on active offshore oil/gas platforms (AOOP) is of importance for effective ocean management and the comprehensive study on marine carbon emissions. In recent years, the methodological framework for AOOP extraction using low-light remote sensing imagery has emerged as a novel and promising approach. Consequently, in this study, we utilized high-resolution low-light images obtained from the Glimmer Imager (GLI) payload aboard the Sustainable Development Goals science Satellite 1 (SDGSAT-1) to facilitate the automatic extraction of AOOP. We applied a time-series low-light remote sensing (TLLRS) method for automatic extraction of AOOP in the northern Gulf of Mexico for the year 2022. The results indicate that the TLLRS method achieves a remarkable precision rate of 95.31% in the automatic extraction of AOOP, with an overall accuracy of 84.92%. Our findings demonstrate the feasibility and efficacy of applying SDGSAT-1 GLI data to monitor AOOP at large-scale. © 2024 SPIE.

关键词： Remote sensing

来源：评论

学校读者我要写书评

暂无评论

Deploying DAPHNE Computational Intelligence on EuroHPC Vega for Benchmarking Randomised Optimisation Algorithms

Deploying DAPHNE Computational Intelligence on EuroHPC Vega ...

引用

International Conference on Broadband Communications for Next Generation Networks and Multimedia Applications (CoBCom)

作者： Aleš Zamuda Mark Dokter Faculty of Electrical Engineering and Computer Science University of Maribor Maribor Slovenia KNOW-CENTER GmbH Research Center for Data-Driven Business & Big Data Analytics Graz Austria

ISBN: (数字)9798350388459

ISBN: (纸本)9798350388466

This paper presents a deployment of DAPHNE (Integrated data Analysis Pipelines for Large-Scale data Management, HPC and Machine Learning) for Computational Intelligence (CI). As CI progresses, e.g. through effort in human societies like the IEEE Computational Intelligence Society (CIS), the latest advances have become more widely popularized. Likewise, with the accessibility of supercomputing resources to advance CI, like with infrastructures such as EuroHPC machines including Vega, this paper demonstrates deployment of Randomised Optimisation Algorithms (ROAs) as a subset of CI algorithms, on EuroHPC Vega. As the Vega is a cluster of distributed nodes, a successful computation on several nodes at once in parallel and concurrently, requires an orchestrated deployment across the nodes, batched using Slurm in this paper. An example ROA benchmarking scenario is then presented in this scope.

关键词： data analysis Pipelines Clustering algorithms Text summarization Machine learning Benchmark testing Timing

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：