检索结果-内蒙古大学图书馆

2024 IEEE International Conference on Systems, Man, and Cybernetics, SMC 2024

作者： Liang, Xuehui Wang, Ruomei Lin, Ge Feng, Jiawei Luo, Yuanmao School of Computer Science and Engineering Sun Yat-sen University National Engineering Research Center of Digital Life Guangzhou China School of Software Engineering Sun Yat-sen University Zhuhai China

ISBN: (纸本)9781665410205

Joint video moment retrieval and highlight detection is a video understanding task that requires the model to construct multimodal interaction between heterogeneous features. Recent Transformer-based models mainly focus on promoting global interaction between features. However, local interaction and temporal asynchronism modeling are not deeply considered. To solve this problem, this paper proposes a dual-branch complementary multimodal interaction mechanism (DCMI), which consists of a global difference feature activation module (GDFA) and a local information dynamic aggregation module (LIDA). GDFA measures the difference between the target element and the global features, thus activating important information. LIDA designs a multimodal heterogeneous graph and constructs asynchronous interaction between heterogeneous features to dynamically aggregate local information. DCMI adaptively fuses the complementary dual branches to improve the model's cognitive and decision-making abilities of global and local information. Comprehensive comparisons with existing methods on public datasets verify the superiority of the proposed model. Extensive ablation experiments and qualitative analysis show the effectiveness and rationality of DCMI, which can promote the interaction between multimodal features. © 2024 IEEE.

关键词： Video analysis

来源：评论

学校读者我要写书评

暂无评论

An Empirical Analysis on Spatial Reasoning Capabilities of Large Multimodal Models

An Empirical Analysis on Spatial Reasoning Capabilities of L...

引用

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Shiri, Fatemeh Guo, Xiao-Yu Far, Mona Golestan Yu, Xin Haffari, Gholamreza Li, Yuan-Fang Department of Data Science & AI Monash University Australia Australian Institute for Machine Learning University of Adelaide Australia School of Electrical Engineering and Computer Science University of Queensland Australia

ISBN: (纸本)9798891761643

Large Multimodal Models (LMMs) have achieved strong performance across a range of vision and language tasks. However, their spatial reasoning capabilities are under-investigated. In this paper, we construct a novel VQA dataset, Spatial-MM, to comprehensively study LMMs' spatial understanding and reasoning capabilities. Our analyses on object-relationship and multi-hop reasoning reveal several important findings. Firstly, bounding boxes and scene graphs, even synthetic ones, can significantly enhance LMMs' spatial reasoning. Secondly, LMMs struggle more with questions posed from the human perspective than the camera perspective about the image. Thirdly, chain of thought (CoT) prompting does not improve model performance on complex multi-hop questions involving spatial relations. Lastly, our perturbation analysis on GQA-spatial reveals that LMMs are much stronger at basic object detection than complex spatial reasoning. We believe our new benchmark dataset and in-depth analyses can spark further research on LMMs spatial reasoning. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Tropical Cyclone Intensity Prediction Using Deep Learning Techniques- A Survey 4

Tropical Cyclone Intensity Prediction Using Deep Learning Te...

引用

4th IEEE International Conference on Applied Electromagnetics, Signal Processing, and Communication, AESPC 2024

作者： Ilayaraja, V. Rajamohana, S.P. Jagadeesan, S. Aishwarya, V. Rakesh, M. School of Computer Science and Engineering Vellore Institute of Technology Tamilnadu Vellore632 014 India School of Engineering & Technology Pondicherry University Department of Computer Science Puducherry Karaikal India School of Computer Science Engineering and Information Systems Vellore Institute of Technology Department of Software and Systems Engineering Tamil Nadu Vellore632 014 India

ISBN: (纸本)9798331528171

Tropical cyclones, characterized by strong winds and heavy rainfall, threaten human life in coastal regions crucial to the economy, including fisheries, agriculture, tourism, and infrastructure. Their frequent occurrence impacts communities reliant on these industries for livelihoods. Accurate estimation of tropical cyclone intensity is vital for disaster preparedness, risk assessment, and timely evacuations. Recent advancements in machine learning and deep learning have been applied to predict cyclone intensity from satellite images, providing insights into cyclone dynamics and enhancing disaster response. This paper analyzes recent research on intensity estimation using various machine learning algorithms and discusses future prospects for improving accuracy and reliability. © 2024 IEEE.

关键词： Tropical cyclone

来源：评论

学校读者我要写书评

暂无评论

Adaptive Evolutionary Reinforcement Learning Algorithm with Early Termination Strategy 23

Adaptive Evolutionary Reinforcement Learning Algorithm with ...

引用

23rd International Conference on Autonomous Agents and Multiagent Systems, AAMAS 2024

作者： Wu, Xiaoqiang Zhu, Qingling Lin, Qiuzhen Chen, Weineng Li, Jianqiang College of Computer Science and Software Engineering Shenzhen University Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China School of Computer Science and Engineering South China University of Technology Guangzhou China

Evolutionary reinforcement learning algorithms (ERLs), which combine evolutionary algorithms (EAs) with reinforcement learning (RL), have demonstrated significant success in enhancing RL performance. However, most ERLs rely heavily on Gaussian mutation operators to generate new individuals. When the standard deviation is too large or small, this approach will result in the production of poor or highly similar offspring. Such outcomes can be detrimental to the learning process of the RL agent, as too many poor or similar experiences are generated by these individuals. In order to alleviate these issues, this paper proposes an Adaptive Evolutionary Reinforcement Learning (AERL) method that adaptively adjusts both the standard deviation and the evaluation process. By tracking the performance of new individuals, AERL maintains the mutation strength within a suitable range without the need for additional gradient computations. Moreover, the proposed AERL approach early terminates unnecessary evaluations and discards experiences arising from poor individuals, thereby resulting in enhanced learning efficiency. Empirical assessments conducted on a variety of continuous control problems demonstrate the effectiveness of the AERL method. © 2024 International Foundation for Autonomous Agents and Multiagent Systems.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Image Compression and Stable Reconstruction Based on Multi-Level Hybrid Feature Guidance 2

Image Compression and Stable Reconstruction Based on Multi-L...

引用

2nd International Conference on Machine Learning, Cloud Computing, and Intelligent Mining, MLCCIM 2023

作者： Zeng, Heng Gong, Nanxu College of Software Engineering Sichuan University Chengdu China School of Computer Science Sichuan University Chengdu China

ISBN: (纸本)9798350328530

Image compression is a technique that reduces storage space and transmission bandwidth requirements by reducing the amount of image data, aiming to improve storage and transmission efficiency while maintaining acceptable image quality. In recent years, learning-based image compression methods have demonstrated superior rate-distortion performance compared to traditional image compression standards. Nevertheless, existing approaches often struggle to simultaneously achieve high compression rates and high-quality reconstruction. To tackle this issue, we present an image compression and stable reconstruction method based on multi-level hybrid feature guidance. Firstly, a hybrid attention module that combines spatial, channel, and shuffle attention is designed to improve the model's ability to capture local features of the image. Secondly, we utilize boundary prediction and frequency domain feature prediction auxiliary branches to guide the network for image compression and reconstruction. Moreover, we propose an image compression and reconstruction algorithm framework based on multi-level feature guidance. Based on experimental results, our method surpasses existing approaches in achieving superior rate-distortion performance. These research results are of great significance to the further development of image compression and reconstruction technology. © 2023 IEEE.

关键词： Image compression

来源：评论

学校读者我要写书评

暂无评论

Modeling and Discovering Data Race with Concurrent Code Property Graphs 23

Modeling and Discovering Data Race with Concurrent Code Prop...

引用

23rd IEEE International Conference on software Quality, Reliability, and Security Companion, QRS-C 2023

作者： Cao, Yang Dong, Yunwei School of Software Northwestern Polytechnical University Xi'an China School of Computer Science and Engineering Northwestern Polytechnical University Xi'an China

ISBN: (纸本)9798350359398

Concurrent programs have brought about higher resource utilization and faster response times. However, the non-determinism in concurrent programs introduces concurrency bugs that are harder to detect than traditional program bugs. These concurrency bugs pose serious threats to the safety and security of software systems. Data races are the most common type of concurrency bug. Previous static analysis techniques either lack precision or face difficulties when applied to large-scale programs. The vulnerability detection tool Joern provides a new static approach that uses a novel source code representation, called code property graph, and effectively mines various types of source code vulnerabilities in large-scale programs through graph traversal. However, because the code property graph does not model the concurrency relationships between threads, Joern does not support the detection of concurrency bugs. Therefore, this paper proposes a property graph-based data race detection method called RaceQuery. This method introduces the concept of property graphs to represent concurrent programs. It constructs the concurrent code property graph based on the control flow graph and enhances it with synchronization operations analysis to improve detection precision. Subsequently, RaceQuery generates a set of concurrent thread pairs for the program based on the concurrent code property graph and models data races through graph traversal. This enables RaceQuery to effectively locate data races in the concurrent code property graph. The method is implemented as a prototype system using the popular graph database Neo4j and successfully detects 2 data races in an open-source project. © 2023 IEEE.

关键词： Static analysis

来源：评论

学校读者我要写书评

暂无评论

LogCAE: An Approach for Log-based Anomaly Detection with Active Learning and Contrastive Learning 35

LogCAE: An Approach for Log-based Anomaly Detection with Act...

引用

35th IEEE International Symposium on software Reliability engineering, ISSRE 2024

作者： Xiao, Pei Jia, Tong Duan, Chiming Cai, Huaqian Li, Ying Huang, Gang Peking University School of Software and Microelectronics Beijing China Peking University Institute for Artificial Intelligence Beijing China Peking University School of Computer Science Beijing China Peking University National Engineering Research Center for Software Engineering Beijing China

ISBN: (纸本)9798350353884

Log-based anomaly detection plays a crucial role in maintaining the reliability of software systems. Unsupervised models are more suitable for real-world usage because they do not rely on huge data labeling efforts. However, their effectiveness is limited because of the lack of supervision of data labels. To balance model effectiveness and labeling efforts, existing approaches enhance model capabilities by incorporating relatively few but key human labels as a golden signal, thereby improving the model ability with acceptable labeling efforts. However, these methods still face limitations of complex human labels and insufficient utilization of human knowledge. In this paper, we introduce LogCAE, a two-stage log anomaly detection approach based on active learning and contrastive learning. It utilizes an unsupervised model to learn from unlabeled log data without human labels and incorporates human knowledge through active learning during online optimization. We employ contrastive learning to optimize the representation of log samples in feature space for more efficient usage of human labels. We conducted experiments on three distinct public log datasets (Thunderbird, BGL, and Zookeeper). The results show that our method improves 12.93% F1-score on average with 6.06% labeled data samples. Besides, our approach is more effective in utilizing human labels than state-of-The-Art approaches. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

SA-ResNet:An Intrusion Detection Method Based on Spatial Attention Mechanism and Residual Neural Network Fusion

引用

computers, Materials & Continua 2025年第5期83卷 3335-3350页

作者： Zengyu Cai Yuming Dai Jianwei Zhang Yuan Feng School of Computer Science and Technology Zhengzhou University of Light IndustryZhengzhou450066China College of Software Engineering Zhengzhou University of Light IndustryZhengzhou450066China Faculty of Information Engineering Xuchang Vocational Technical CollegeXuchang461000China School of Electronic Information Zhengzhou University of Light IndustryZhengzhou450066China

The rapid development and widespread adoption of Internet technology have significantly increased Internet traffic,highlighting the growing importance of network *** Detection Systems(IDS)are essential for safeguarding network *** address the low accuracy of existing intrusion detection models in identifying network attacks,this paper proposes an intrusion detection method based on the fusion of Spatial Attention mechanism and Residual Neural Network(SA-ResNet).Utilizing residual connections can effectively capture local features in the data;by introducing a spatial attention mechanism,the global dependency relationships of intrusion features can be extracted,enhancing the intrusion recognition model’s focus on the global features of intrusions,and effectively improving the accuracy of intrusion *** proposed model in this paper was experimentally verified on theNSL-KDD *** experimental results showthat the intrusion recognition accuracy of the intrusion detection method based on SA-ResNet has reached 99.86%,and its overall accuracy is 0.41% higher than that of traditional Convolutional Neural Network(CNN)models.

关键词： Intrusion detection deep learning residual neural network spatial attention mechanism

来源：评论

学校读者我要写书评

暂无评论

CAPS: An Efficient Whole-Program Critical Paths Search Framework for Large-Scale software 35

CAPS: An Efficient Whole-Program Critical Paths Search Frame...

引用

35th International Conference on software engineering and Knowledge engineering, SEKE 2023

作者： Li, Peiyang Liu, Zixin Su, Yuening Wang, Hao Jiang, Bo State Key Laboratory of Software Development Environment School of Computer Science and Engineering Beihang University China

Tracking the flow of external inputs in a program with taint-analysis techniques can help developers better identify potential security vulnerabilities in the software. However, directly using the static taint analysis provided by Clang Static Analyzer is inefficient for large-scale software due to the huge but redundant ExplodedGraph generated. Therefore, we propose an efficient Whole-Program Critical Paths Search (CAPS) framework. It first performs a set of optimizations to reduce the ExplodedGraph of each function. Then, it constructs a global exploded graph by inserting call edges among the reduced ExplodedGraphs for each function within the Neo4j graph database. Finally, it proposes loop removal and graph segmentation to optimize the search process for critical paths on the global exploded graph. Our experiments on 3 large-scale software show that CAPS can significantly improve the efficiency of critical path search for large-scale software. © 2023 Knowledge Systems Institute Graduate school. All rights reserved.

关键词： Static analysis

来源：评论

学校读者我要写书评

暂无评论

Fast Verification of Network Configuration Updates

引用

computers, Materials & Continua 2023年第1期74卷 293-311页

作者： Jiangyuan Yao Zheng Jiang Kaiwen Zou Shuhua Weng Yaxin Li Deshun Li Yahui Li Xingcan Cao Scool of Computer Science and Technology Hainan UniversityHaikou570228China School of Cyberspace Security Hainan UniversityHaikou570228China School of Software Engineering Jilin UniversityChangchun130012China School of Software Beijing Jiaotong UniversityBeijing100091China University of British Columbia VancouverV5K1K5Canada

With the expansion of network services,large-scale networks have progressively become *** network status changes rapidly in response to customer needs and configuration changes,so network configuration changes are also very ***,no matter what changes,the network must ensure the correct conditions,such as isolating tenants from each other or guaranteeing essential *** changes occur,it is necessary to verify the after-changed ***,for the verification of large-scale network configuration changes,many current verifiers show poor *** order to solve the problem ofmultiple global verifications caused by frequent updates of local configurations in large networks,we present a fast configuration updates verification tool,FastCUV,for distributed control *** aims to enhance the efficiency of distributed control plane verification for medium and large networks while ensuring *** paper presents a method to determine the network range affected by the configuration *** present a flow model and graph structure to facilitate the design of verification algorithms and speed up *** scheme verifies the network area affected by obtaining the change of the Forwarding Information Base(FIB)before and *** supports rich network attributes,meanwhile,has high efficiency and correctness *** experimental verification and result analysis,our method outperforms the state-of-the-art method to a certain extent.

关键词： Network verification configuration updates network control plane forwarding information base

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：