检索结果-内蒙古大学图书馆

Flexible ViG: Learning the Self-Saliency for Flexible Object Recognition

IEEE Transactions on Circuits and Systems for Video Technology 2025年

作者： Yang, Kunshan Zuo, Lin Jing, Mengmeng Tian, Xianlong He, Kunbin Ding, Yongqi University of Electronic Science and Technology of China School of Information and Software Engineering ChengDu611731 China University of Electronic Science and Technology of China School of Computer Science and Engineering ChengDu611731 China

Existing computer vision methods mainly focus on the recognition of rigid objects, whereas the recognition of flexible objects remains unexplored. Recognizing flexible objects poses significant challenges due to their inherently diverse shapes and sizes, translucent attributes, ambiguous boundaries, and subtle inter-class differences. In this paper, we claim that these problems primarily arise from the lack of object saliency. To this end, we propose the Flexible Vision Graph Neural Network (FViG) to optimize the self-saliency and thereby improve the discrimination of the representations for flexible objects. Specifically, on one hand, we propose to maximize the channel-aware saliency by extracting the weight of neighboring graph nodes, which is employed to identify flexible objects with minimal inter-class differences. On the other hand, we maximize the spatial-aware saliency based on clustering to aggregate neighborhood information for the centroid graph nodes. This introduces local context information and enables extracting of consistent representation, effectively adapting to the shape and size variations in flexible objects. To verify the performance of flexible objects recognition thoroughly, for the first time we propose the Flexible Dataset (FDA), which consists of various images of flexible objects collected from real-world scenarios or online. Extensive experiments evaluated on our FDA, FireNet, CIFAR-100 and ImageNet-Hard datasets demonstrate the effectiveness of our method on enhancing the discrimination of flexible objects. © 1991-2012 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

CAC 2.0: A Corrupt and Correct Logic Locking Technique Resilient to Structural Analysis Attacks

arXiv

引用

arXiv 2024年

作者： Aksoy, Levent Yasin, Muhammad Pagliarini, Samuel Department of Computer Systems Tallinn University of Technology Tallinn Estonia Department of Computer and Software Engineering National University of Sciences and Technology Islamabad Pakistan Department of Electrical and Computer Engineering Carnegie Mellon University Pittsburgh United States

Logic locking proposed to protect integrated circuits from serious hardware threats has been studied extensively over a decade. In these years, many efficient logic locking techniques have been proven to be broken. The state-of-the-art logic locking techniques, including the prominent corrupt and correct (CAC) technique, are resilient to satisfiability (SAT)-based and removal attacks, but vulnerable to structural analysis attacks. To overcome this drawback, this paper introduces an improved version of CAC, called CAC 2.0, which increases the search space of structural analysis attacks using obfuscation. To do so, CAC 2.0 locks the original circuit twice, one after another, on different nodes with different number of protected primary inputs using CAC, while hiding original protected primary inputs among decoy primary inputs. This paper also introduces an open source logic locking tool, called HIID, equipped with well-known techniques including CAC 2.0. Our experiments show that CAC 2.0 is resilient to existing SAT-based, removal, and structural analysis attacks. To achieve this, it increases the number of key inputs at most 4× and the gate-level area between 30.2% and 0.8% on circuits with low and high complexity with respect to CAC. © 2024, CC BY.

关键词： Structural analysis

来源：评论

学校读者我要写书评

暂无评论

Research on Obstacle Avoidance Algorithm of UAVs Reinforcement Learning based on Vision

Research on Obstacle Avoidance Algorithm of UAVs Reinforceme...

引用

Frontiers Technology of Information and computer (ICFTIC), IEEE International Conference on

作者： Yuan Gao Tianwei Shi Teng Xu Jianbang Ding Rongchen Jing School of Computer Science and Software Engineering University of Science and Technology Liaoning Anshan China

In recent years, the combination of deep reinforcement learning and unmanned aerial vehicle (UAV) to achieve autonomous flight has been a hot research field. In this paper, an obstacle avoidance navigation algorithm (PA-SAC) based on priority experience buffer pool (PEBP), attention, and Soft-Actor-Critic (SAC) is proposed to solve the continuous space obstacle avoidance navigation problem of UAVs by using a deep reinforcement learning algorithm. In the two simulation experiments, the success rate of the PA-SAC algorithm was 95.6% in the known environment and 73% in the unknown environment. These results demonstrate that the PA-SAC algorithm can achieve autonomous obstacle avoidance for UAVs with deep image input.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SRGAN Optimization Through Enhanced Network Structure and Multi-Loss Functions

SRGAN Optimization Through Enhanced Network Structure and Mu...

引用

Data Science and computer Application (ICDSCA), IEEE International Conference on

作者： Zexin Chen Xiaoxin Chen College of Computer Science and Software Engineering Shenzhen University Shenzhen China College of Medical Information Engineering Guangdong Pharmaceutical University Guangzhou China

ISBN: (数字)9798350368239

ISBN: (纸本)9798350368246

An enhanced SRGAN image super-resolution algorithm is proposed to address SRGAN's training instability and limited global information capture. This enhancement optimizes both the network structure and loss function. In network structure optimization, group normalization is introduced as an alternative to batch normalization, addressing its dependency on batch size and improving stability during small-batch training. Additionally, Global Context Block (GCBlock) is integrated into the residual block, enhancing the network's global information capture and detail representation capabilities. In regard to the loss function, this paper introduces L1 loss and MS-SSIM loss as alternatives to the conventional MSE loss. This approach aims to enhance the quality of the generated image in both pixel space and structural information while reducing the smoothing effect. In the experimental section, the ablation experiments evaluate the performance of models with and without GCBlock on five test sets, including U rban1 00 and Manga109. The results demonstrate that GCBlock improves both detail expressiveness and global information capture, with an average PSNR gain of 0.1388 and SSIM improvement of 0.008. In comparison experiments, the improved algorithm achieved an average PSNR increase of 0.270 and an average SSIM increase of 0.018 on the aforementioned test set, indicating a significant improvement in detail expression and visual quality. Finally, the limitations of the enhanced algorithm are discussed, and potential future research directions are proposed.

关键词： Training Visualization Smoothing methods Computational modeling Superresolution Stability analysis Data models Optimization Image reconstruction Thermal stability

来源：评论

学校读者我要写书评

暂无评论

Fragmentation-Aware VNF Placement: A Deep Reinforcement Learning Approach

Fragmentation-Aware VNF Placement: A Deep Reinforcement Lear...

引用

IEEE International Conference on Communications (ICC)

作者： Ramy Mohamed Marios Avgeris Aris Leivadeas Ioannis Lambadaris Department of Systems and Computer Engineering Carleton University Ottawa Canada Department of Software and IT Engineering École de technologie sup érieure Montreal Canada

ISBN: (数字)9781728190549

ISBN: (纸本)9781728190556

In this paper we address the challenge of efficiently deploying Virtual Network Functions (VNFs) in network infrastructures. This is particularly crucial when facing resource fragmentation, where available resources are not fully utilized due to the fluctuating allocation and deallocation of virtual network requests. Traditional optimization techniques often fall short in managing the dynamic complexities of VNF placement. To overcome this, we introduce a novel online VNF placement strategy using Deep Reinforcement Learning (DRL) combined with a Reward Constrained Policy Optimization (RCPO). This method leverages the flexibility of DRL and the constraint integration capacity of RCPO, ensuring compliance with performance and resource limitations while minimizing resource fragmentation. The results demonstrate that our DRL-based method surpasses existing methods, resulting in more effective resource management and less resource fragmentation.

关键词： Metalearning Adaptive learning Costs Transfer learning Complex networks Deep reinforcement learning Complexity theory

来源：评论

学校读者我要写书评

暂无评论

Prompting Continual Person Search 24

Prompting Continual Person Search

引用

32nd ACM International Conference on Multimedia, MM 2024

作者： Zhang, Pengcheng Yu, Xiaohan Bai, Xiao Zheng, Jin Ning, Xin School of Computer Science and Engineering State Key Laboratory of Complex & Critical Software Environment Jiangxi Research Institute Beihang University Beijing China School of Computing Macquarie University Sydney Australia Institute of Semiconductors Chinese Academy of Sciences Beijing China

ISBN: (纸本)9798400706868

The development of person search techniques has been greatly promoted in recent years for its superior practicality and challenging goals. Despite their significant progress, existing person search models still lack the ability to continually learn from increasing real-world data and adaptively process input from different domains. To this end, this work introduces the continual person search task that sequentially learns on multiple domains and then performs person search on all seen domains. This requires balancing the stability and plasticity of the model to continually learn new knowledge without catastrophic forgetting. For this, we propose a Prompt-based Continual Person Search (PoPS) model in this paper. First, we design a compositional person search transformer to construct an effective pre-trained transformer without exhaustive pre-training from scratch on large-scale person search data. This serves as the fundamental for prompt-based continual learning. On top of that, we design a domain incremental prompt pool with a diverse attribute matching module. For each domain, we independently learn a set of prompts to encode the domain-oriented knowledge. Meanwhile, we jointly learn a group of diverse attribute projections and prototype embeddings to capture discriminative domain attributes. By matching an input image with the learned attributes across domains, the learned prompts can be properly selected for model inference. Extensive experiments are conducted to validate the proposed method for continual person search. The source code is available at https://***/PatrickZad/PoPS. © 2024 ACM.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Based Anomaly Detection in 5G Cellular Networks

Deep Learning-Based Anomaly Detection in 5G Cellular Network...

引用

International Conference on Information, Intelligence, Systems and Applications (IISA)

作者： Umar Farooq Aroosa Hameed Aris Leivadeas Ioannis Lambadaris Department of Software and IT Engineering École de technologie supérieure Montreal Canada Department of Systems and Computer Engineering Carleton University Ottawa Canada

ISBN: (数字)9798350368833

ISBN: (纸本)9798350368840

The telecommunication industry saw a dramatic shift with the introduction of 5G technologies. This new cellular generation brought lightening speed, massive capacity, and better connectivity. Despite the many introduced opportunities, several challenges emerge at the same time, especially in the field of performance assurance. An example of such challenges is the rapid identification of network anomalies, which is critical for maintaining network performance and ensuring user satisfaction. Traditional techniques of detecting anomalies have fallen short given the unique operating requirements of 5G networks. To address this issue, this paper presents a novel technique applying deep learning for the detection of cellular network anomalies. Specifically, we leverage and exploit the capabilities of several Long-Short-Term-Memory (LSTM) and Artificial Neural Networks (ANN) flavors, such as LSTM with ANN and Bidirection-LSTM (BiLSTM) with ANN, to excel at identifying network anomalies for the preservation and efficiency of 5G networks. Our work includes extensive experimentation and the attained results show that our proposed models are overwhelmingly adapting to preserving the integrity of the network in the fast-paced, ever-changing realm of cellular networks.

关键词： Cellular networks Industries Deep learning Adaptation models Accuracy 5G mobile communication Quality of service Communications technology Sustainable development Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Prediction of Air Quality Index Using Random Forest and Prophet Tool

Prediction of Air Quality Index Using Random Forest and Prop...

引用

IEEE International Conference on System of Systems engineering (SoSE)

作者： Abhishek Walia Ajay Paliwal Sanjay Patidar Rakeshkumar Mahto Software Engineering Department Delhi Technological University Delhi India Department of Electrical and Computer Engineering California State University Fullerton California USA

ISBN: (数字)9798350365917

ISBN: (纸本)9798350365924

The Air Quality Index (AQI) serves as a critical tool for informing the public about the negative impact of air pollution on human health. In Indian cities, where air pollution levels have been rising sharply, accurate assessment and forecasting of AQI are more important than ever. This study evaluates urban air quality and forecasts future AQI by analyzing a wide array of pollutants, including nitrogen oxides (NO x ), ammonia, carbon monoxide, sulfur dioxide, ozone, benzene, toluene, PM2.5, PM10, nitric oxide, nitrogen dioxide, and xylene. We employ a Random Forest classifier for detection and use the Prophet library for forecasting. The performance of our model is demonstrated by precision, accuracy, and recall metrics, with the model achieving an accuracy score of 99.57%. This high level of accuracy indicates the model’s robust capability in predicting AQI and underscores its potential utility in urban air quality management.

关键词： Accuracy Atmospheric modeling Urban areas Air pollution Indexes Nitrogen Sulfur

来源：评论

学校读者我要写书评

暂无评论

Building a Full-text Search Index Using "Transformer" Neural Network

Building a Full-text Search Index Using "Transformer" Neural...

引用

International Conference on Application of Information and Communication Technologies (AICT)

作者： Vyacheslav Dobrynin Roman Abramovich Alexey Platonov Faculty of Software Engineering and Computer Systems ITMO University Saint Petersburg Russian Federation

The use of deep neural networks in information retrieval significantly improves its effectiveness, but negatively affects the performance of the process. To deal with this, we propose a new ranking model that uses the deep neural network of the "Transformer" architecture (in particular, BERT) for efficient information retrieval. In accordance with the proposed approach, contextualized vector representations are extracted from documents during indexing, after which these representations are clustered for each independent token. The resulting clusters reflect different meanings of the words and are indirectly used as inverted index keys. The values represent the documents in which these contextualized word meanings occur, along with the distances from each document to the contextualized embedding. Thus, after the indexing process, we obtain an index containing pre-calculated distances between the contextualized meanings of dictionary elements and documents. This approach helps us avoid the performance overhead of calculating distances online. At the search stage, the query is transformed into a set of contextualized vectors representing each query token, which allows us to use these vectors to retrieve most semantically close neighbor-tokens and use them to extract relevant documents from the index. This way of searching for contextualized embeddings consumes less memory and is more performant due to the use of an inverted index.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Federated Social Recommendation with Rényi Differential Privacy

Federated Social Recommendation with Rényi Differential Pri...

引用

computer Technology and Information Science (ISCTIS), International Symposium on

作者： Tian jiao Wang Jie Wu School of Computer and Software Engineering University of Science and Technology Liaoning Anshan Liaoning China

A social recommendation system based on graph neural networks is a system that uses social relationships between users to generate personalized recommendations. To improve recommendation accuracy, it is usually necessary to collect a large amount of user behavior data, which may lead to user privacy breaches and data misuse. Existing privacy protection methods often sacrifice recommendation effectiveness or increase computational costs. Therefore, this article proposes a federated social recommendation system (FRSRec) based on Rényi differential privacy. By combining hash encryption technology with Rényi differential privacy, user data privacy can be protected while ensuring recommendation effectiveness and computational efficiency. This article conducted experimental evaluations on three real datasets. The experimental results show that the FRSRec model achieves optimal results in both MAE and RMSE, demonstrating that the FRSRec model can effectively achieve privacy protection while improving recommendation quality.

关键词：

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：