检索结果-内蒙古大学图书馆

作者： Ma, Chenlin Yang, Hao Wu, Shangyu Wang, Yi Mao, Rui Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China City University of Hong Kong Department of Computer Science Hong Kong Shenzhen University College of Computer Science and Software Engineering The Shenzhen Institute of Computing Sciences Shenzhen518060 China

The log-structured merge-tree (LSM-tree)-based key-value store has been widely adopted by many large-scale data storage applications for its excellent write performance. However, such write performance gains mainly come from scarifying read performance due to the leveled and log-structured intrinsic characteristics of the LSM-tree. Therefore, the critical challenge of the existing LSM-tree is how to improve the read efficiency by reducing read amplification. This article for the first time proposes Tidal-tree-Mem, a novel data structure where data flow inside the LSM-tree-like Tidal waves. First, a floating strategy is proposed to allow frequently accessed files at the bottom of the LSM-tree to move to higher positions, reducing read amplification. Second, a stretching strategy is proposed to vary the shape of the LSM-tree to adapt to workloads with different characteristics. To evaluate the performance of Tidal-tree-Mem, we conduct a series of experiments using standard benchmarks from YCSB. The experimental results show that Tidal-tree-Mem can effectively reduce read amplification and the overall latency by over 71.94% and 47.34%, respectively, compared with representative schemes. © 1982-2012 IEEE.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Preprocessing Method for Color-Adaptive Pix2Pix Generative Neural Network Data

引用

Journal of Institute of Control, Robotics and Systems 2023年第6期29卷 521-526页

作者： Lee, Hayoung Son, Bong-Ki Lee, Jaeho Department of IT Media Engineering Duksung Women’s University Korea Republic of Department of Computer Science and Engineering Seowon University Korea Republic of Department of Software Duksung Women’s University Korea Republic of

Adversarial generative neural networks (GANs) have received increased research interest in recent years, and they have been applied in various fields, such as image generation, transformation, and restoration. Pix2Pix, a popular GAN model, trains data in which input and output data are paired. However, because this method is based on sketches, it cannot consider colors once the training is completed. Not only to apply colors requires coloring on edges of all training data, which requires high effort to build data, but also it is difficult to automatically generate edges. In this paper, we proposed a method that considers colors and performs the automatic generation of edges required for Pix2Pix. The proposed method extracts edge data by using the canny edge detection algorithm and constructs color edge and color background data by using the extracted edge data. The applicability of the proposed method was verified by conducting experiments in the fashion field. © ICROS 2023.

关键词： Color

来源：评论

学校读者我要写书评

暂无评论

Parameter-Efficient Federated Cooperative Learning for 3D Object Detection in Autonomous Driving

引用

IEEE Internet of Things Journal 2025年第12期12卷 20314-20325页

作者： Chi, Fangyuan Wang, Yixiao Nasiopoulos, Panos Leung, Victor C.M. University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T 1Z4 Canada Toronto AI Lab LG Electronics TorontoONM5V 1M3 Canada Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China

In the rapidly evolving field of autonomous driving, accurately detecting and understanding dynamic environments remains a challenge. Federated learning (FL) offers a promising approach by integrating decentralized models from multiple Connected Autonomous Vehicles (CAVs) to enhance the performance of Deep Learning (DL) based object detection methods. However, traditional FL faces hurdles such as extensive data synchronization requirements, limited data variance, and high communication costs. This paper introduces a federated cooperative learning framework that addresses these challenges by combining data from both CAVs and Roadside Units (RSUs). The framework combines local cooperative perception with global federated learning through a parameter-efficient federated learning adapter (PEFLA) and a lazy communication (LazyComm) strategy, improving DL-based object detection capabilities across diverse driving scenarios while significantly reducing bandwidth requirements. We also present a novel multi-agent-multi-town dataset V2X-Fed, specifically developed to validate the effectiveness of our approach under various conditions. Notably, our framework retains 97.51% of the detection accuracy achieved by full-model federated learning, while utilizing only 1.4% of the bandwidth typically required, demonstrating substantial improvements over conventional federated learning strategies. This study underscores the potential of our tailored approach to substantially enhance autonomous vehicle technologies with minimal resource utilization. © 2014 IEEE.

关键词： Bandwidth

来源：评论

学校读者我要写书评

暂无评论

Adapt PointFormer: 3D Point Cloud Analysis via Adapting 2D Visual Transformers 27

Adapt PointFormer: 3D Point Cloud Analysis via Adapting 2D V...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Li, Mengke Li, Da Yang, Guoqing Cheung, Yiu-Ming Huang, Hui Shenzhen China VCC College of Computer Science and Software Engineering Shenzhen University Shenzhen China Department of Computer Science Hong Kong Baptist University Hong Kong

ISBN: (纸本)9781643685489

Pre-trained large-scale models have exhibited remarkable efficacy in computer vision, particularly for 2D image analysis. However, when it comes to 3D point clouds, the constrained accessibility of data, in contrast to the vast repositories of images, poses a challenge for the development of 3D pre-trained models. This paper therefore attempts to directly leverage pre-trained models with 2D prior knowledge to accomplish the tasks for 3D point cloud analysis. Accordingly, we propose the Adaptive PointFormer (APF), which fine-tunes pre-trained 2D models with only a modest number of parameters to directly process point clouds, obviating the need for mapping to images. Specifically, we convert raw point clouds into point embeddings for aligning dimensions with image tokens. Given the inherent disorder in point clouds, in contrast to the structured nature of images, we then sequence the point embeddings to optimize the utilization of 2D attention priors. To calibrate attention across 3D and 2D domains and reduce computational overhead, a trainable PointFormer with a limited number of parameters is subsequently concatenated to a frozen pre-trained image model. Extensive experiments on various benchmarks demonstrate the effectiveness of the proposed APF. The source code and more details are available at https://***/research/2024/PointFormer. © 2024 The Authors.

关键词：

来源：评论

学校读者我要写书评

暂无评论

BiEfficient: Bidirectionally Prompting Vision-Language Models for Parameter-Efficient Video Recognition 17th

BiEfficient: Bidirectionally Prompting Vision-Language Model...

引用

17th Asian Conference on computer Vision, ACCV 2024

作者： He, Haichen Liu, Weibin Xing, Weiwei School of Computer Science and Technology Beijing Jiaotong University Beijing China School of Software Engineering Beijing Jiaotong University Beijing China

ISBN: (纸本)9789819609079

Vision-language models (VLMs) pre-trained on large-scale image-text pairs have shown great success in various image tasks. However, how to efficiently transfer such powerful VLMs into video domain is still an open problem. Given that full finetuning VLMs for video tasks could be computationally expensive, recent studies turn their focus on parameter-efficient finetuning (PEFT). The great potential of VLMs lies in leveraging the bidirectional semantic connections between the two modalities of vision and language. Nevertheless, most current PEFT methods use the vision-only framework and usually ignore the semantic connections between vision and language. In this paper, we propose a novel method called BiEfficient, which use bidirectional prompting schemes to efficiently transfer the VLM to video recognition task with a small number of tunable parameters: 1) Vision-to-Language: we propose two prompt mechanisms, Pre-Prompt and Post-Prompt, which act before and after the text encoder respectively to generate discriminative video-level text representation for each input video. 2) Language-to-Vision: we propose Word-Guided Visual-Prompt, which enhances the temporal modeling of videos using textual knowledge in an almost parameter-free manner. Experiments on Kinetics-400, UCF-101, HMDB-51 demonstrate that the proposed method can achieve comparable or even better performance to the full finetuning methods with much fewer tunable parameters across closed-set and zero-shot video recognition benchmarks. Code is available here: https://***/WbLiuBJTULab/BiEfficient. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Discussion on the Influence of Anti-normalization Parameter Selection of BP Neural Network on Mending Seepage Pressure Data of Tailing Dam

Discussion on the Influence of Anti-normalization Parameter ...

引用

2022 International Joint Conference on Information and Communication engineering, JCICE 2022

作者： Zhu, Hongli Jin, Meiqi Wen, Yumeng Lu, Ling Liaoning University of Science and Technology School of Computer and Software Engineering Department of Software Engineering Liaoning Anshan China Liaoning University of Science and Technology School of Computer and Software Engineering Department of Network Engineering Liaoning Anshan China Liaoning University of Science and Technology School of Computer and Software Engineering Liaoning Anshan China

ISBN: (数字)9781665460675

ISBN: (纸本)9781665460675

The height of Infiltration line of tailings dam is a very important parameter for the safety of tailings reservoir. Infiltration line is the intersection line between the free water surface and the cross section of the dam body, that the free water surface is formed by the infiltration of water from the earth dam to the downstream. When the buried depth of the pressure sensors at each monitoring point could be obtained and the groundwater seepage pressure of the tailings dam could be measured too, the height of the free water surface can be calculated then. The commonly used pressure sensor is vibrating wire sensor. Among the many seepage pressure monitoring points, because of the lines, power supply or some other reasons, it often occurs that one or some points do not collect groundwater seepage pressure data for a period of time. For the missing data, BP neural network can be used to mend the missed data. Of course, the patched data should be marked to distinguish it from the real collected data. The mending method of missing seepage pressure data using BP neural network model is analyzed. The influence of anti-normalization parameters on mending data is analyzed. In the process of anti-normalization, the maximum and minimum values of the output data of the training data set are not put to use. Instead, a sequences is formed with the existing output data of some range before and after the mission data segment. The maximum and minimum values of the sequence are selected to use as the anti-normalization parameters. This will make the patched data values closer to the actual collected values. © 2022 IEEE.

关键词： Groundwater

来源：评论

学校读者我要写书评

暂无评论

Toward actionable testing of deep learning models

引用

Science China(Information Sciences) 2023年第7期66卷 300-302页

作者： Yingfei XIONG Yongqiang TIAN Yepang LIU Shing-Chi CHEUNG Key Laboratory of High Confidence Software Technologies(Peking University) Ministry of Education School of Computer Science Peking University Department of Computer Science and Engineering The Hong Kong University of Science and Technology School of Computer Science University of Waterloo Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Department of Computer Science and Engineering Southern University of Science and Technology

Deep learning has become an important computational paradigm in our daily lives with a wide range of applications,from authentication using facial recognition to autonomous driving in smart vehicles. The quality of the deep learning models, i.e., neural architectures with parameters trained over a dataset, is crucial to our daily living and economy.

关键词：

来源：评论

学校读者我要写书评

暂无评论

An Exploratory Study on Code Quality, Testing, Data Accuracy, and Practical Use Cases of IoT Wearables 7

An Exploratory Study on Code Quality, Testing, Data Accuracy...

引用

7th Conference on Cloud and Internet of Things, CIoT 2024

作者： Minani, Jean Baptiste El Fellah, Yahia Ahmed, Sanam Sabir, Fatima Moha, Naouel Gueheneuc, Yann-Gael Concordia University Computer Science and Software Engineering Department MontrealQC Canada École de Technologie Supérieure Département de Génie Logiciel et des Ti MontrealQC Canada University of the Punjab Department of Computer Sciences and Software Engineering Lahore Pakistan

ISBN: (纸本)9798350356717

The growth of the Internet of Things (IoT), particularly in wearable devices like Fitbits, has raised challenges related to source code quality, testing, data accuracy, and practical applications. This paper investigates issues in Fitbit apps by (1) analyzing GitHub repositories of Fitbit projects to identify code quality issues, (2) using Large Language Models (LLMs) to automate testing, (3) comparing data variations across different Fitbit models, and (4) experimenting with real-world use cases for Fitbit devices. Our analysis of 16 GitHub repositories revealed code quality issues in Fitbit apps, highlighting the need for better practices. Using LLMs like ChatGPT-4, we generated unit tests with 100% coverage. Data comparisons across Fitbit Versa models showed consistent accuracy. Finally, we showed the potential of wearable devices in the real-world with two practical use cases: health monitoring with robotic assistance and location-based tracking. These findings open new avenues for research in wearables. © 2024 IEEE.

关键词： Data accuracy

来源：评论

学校读者我要写书评

暂无评论

Local Optimisation Based Detection of Weak Targets in Isolated Forests 24

Local Optimisation Based Detection of Weak Targets in Isolat...

引用

8th International Conference on Electronic Information Technology and computer engineering, EITCE 2024

作者： Hu, Jurong Xing, Yanxiao Li, Yuxiao School of Information Science and Engineering Hohai University Jiangsu Nanjing China School of Computer and Software Hohai University Jiangsu Nanjing China

ISBN: (纸本)9798400710094

Aiming at the problem that the iForest algorithm is not sensitive enough to local anomalies and produces a large number of false alarms in the detection results on some low sea state datasets, this paper proposes the iForest algorithm based on an improved splitting strategy. The algorithm introduces the Gini impurity, incorporates the Gini impurity parameter into the splitting strategy, and is more targeted in the selection of splitting attributes and splitting points, thus better adapting to the problem of detecting weak targets under different sea states, and has an excellent performance of high accuracy, low time duration, and low false alarm rate compared with the traditional methods in anomaly detection. © 2024 Copyright held by the owner/author(s).

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Stage-Specific Reinforcement Learning-Based Firewall for IoT Security Against Okiru Botnet 23

Stage-Specific Reinforcement Learning-Based Firewall for IoT...

引用

23rd International Conference on Cyberworlds, CW 2024

作者： Cho, Hyeon Seok Lee, Jeong Eun Oh, Sang Ho Pukyong National University Department of Computer Engineering & Artificial Intelligence Busan Korea Republic of Busan University of Foreign Studies Department of Computer Software Busan Korea Republic of

ISBN: (纸本)9798331527174

The emergence of sophisticated variant botnet attacks like Okiru poses significant security challenges to the proliferating IoT devices. This study proposes a reinforcement learning (RL)-based defense system for IoT networks. Our system employs an RL agent to analyze network packets in real-time and implement adaptive security measures. By leveraging AI techniques, this research offers a proactive approach to addressing complex cyber threats in IoT environments. © 2024 IEEE.

关键词： Bot (Internet)

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：