检索结果-内蒙古大学图书馆

2024 International Conference on Computer Graphics, Artificial Intelligence, and data Processing, ICCAID 2024

作者： Ma, Liting Zhao, Junzhe Hsiang, Han-Cheng School of Data and Computer Science Xiamen Institute of Technology Xiamen China

ISBN: (纸本)9781510689275

In recent years, detecting objects in aerial images has emerged as a crucial area of study within the domain of computer vision. However, due to obstacles like the limited size of objects, dense distributions, and class imbalance in these images, traditional object detection methods often face issues like missed detections and low performance. In response to these challenges, this study introduces an enhanced YOLOv8s model designed to optimize small object detection. First, to enhance feature extraction capabilities for objects of different scales, the C2f structure within the backbone of the baseline YOLOv8s model is modified by PKI(Poly Kernel Inception)Block from PKINet, thereby strengthening its recognition capabilities across multi-scale targets. Second, to further elevate the efficiency of the neck network for detecting small targets, the MCAttention module is incorporated as part of the C2f design in the neck, leading to the enhanced C2fMCAttn module, which augments the neck network's effectiveness in modeling features from small object areas. Moreover, a 160×160 detection head is added to the detection layer, tailored to enhance the detection of small objects. The experiments were conducted on the VisDrone2019 drone aerial imagery dataset. The results indicate that the improved model outperformed the baseline YOLOv8s model. It achieved a 5.3% increase in precision, a 5% increase in recall and a 4.4% increase in mAP@50. Furthermore, the proposed model outperforms other mainstream object detection methods, exhibiting superior detection effectiveness. © 2025 SPIE.

关键词： MCAttention PKI Block Small object detection YOLOv8s

来源：评论

学校读者我要写书评

暂无评论

Uncertainty for Active Learning on Graphs 41

Uncertainty for Active Learning on Graphs

引用

41st International Conference on Machine Learning, ICML 2024

作者： Fuchsgruber, Dominik Wollschläger, Tom Charpentier, Bertrand Oroz, Antonio Günnemann, Stephan School of Computation Information and Technology Technical University of Munich Germany Munich Data Science Institute Germany

Uncertainty Sampling is an Active Learning strategy that aims to improve the data efficiency of machine learning models by iteratively acquiring labels of data points with the highest uncertainty. While it has proven effective for independent data its applicability to graphs remains under-explored. We propose the first extensive study of Uncertainty Sampling for node classification: (1) We benchmark Uncertainty Sampling beyond predictive uncertainty and highlight a significant performance gap to other Active Learning strategies. (2) We develop ground-truth Bayesian uncertainty estimates in terms of the data generating process and prove their effectiveness in guiding Uncertainty Sampling toward optimal queries. We confirm our results on synthetic data and design an approximate approach that consistently outperforms other uncertainty estimators on real datasets. (3) Based on this analysis, we relate pitfalls in modeling uncertainty to existing methods. Our analysis enables and informs the development of principled uncertainty estimation on graphs. Copyright 2024 by the author(s)

关键词： Active learning

来源：评论

学校读者我要写书评

暂无评论

KubeFuzzer:Automating RESTful API Vulnerability Detection in Kubernetes

引用

Computers, Materials & Continua 2024年第10期81卷 1595-1612页

作者： Tao Zheng Rui Tang Xingshu Chen Changxiang Shen School of Cyber Science and Engineering Sichuan UniversityChengdu610065China Cyber Science Research Institute Sichuan UniversityChengdu610065China Key Laboratory of Data Protection and Intelligent Management(Sichuan University) Ministry of EducationChengdu610065China

RESTful API fuzzing is a promising method for automated vulnerability detection in Kubernetes *** tools struggle with generating lengthy,high-semantic request sequences that can pass Kubernetes API gateway *** address this,we propose KubeFuzzer,a black-box fuzzing tool designed for Kubernetes RESTful *** utilizes Natural Language Processing(NLP)to extract and integrate semantic information from API specifications and response messages,guiding the generation of more effective request *** evaluation of KubeFuzzer on various Kubernetes clusters shows that it improves code coverage by 7.86%to 36.34%,increases the successful response rate by 6.7%to 83.33%,and detects 16.7%to 133.3%more bugs compared to three leading *** identified over 1000 service crashes,which were narrowed down to 7 unique *** tested these bugs on 10 real-world Kubernetes projects,including major providers like AWS(EKS),Microsoft Azure(AKS),and Alibaba Cloud(ACK),and confirmed that these issues could trigger service *** have reported and confirmed these bugs with the Kubernetes community,and they have been addressed.

关键词： Kubernetes RESTful APIs API fuzzing black-box fuzzing

来源：评论

学校读者我要写书评

暂无评论

Few-shot Prompting for Pairwise Ranking: An Effective Non-Parametric Retrieval Model

Few-shot Prompting for Pairwise Ranking: An Effective Non-Pa...

引用

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

作者： Sinhababu, Nilanjan Parry, Andrew Ganguly, Debasis Samanta, Debasis Mitra, Pabitra Centre for Computational and Data Sciences IIT Kharagpur India School of Computing Science University of Glasgow United Kingdom Department of Computer Science and Engineering IIT Kharagpur India

ISBN: (纸本)9798891761681

A supervised ranking model, despite its effectiveness over traditional approaches, usually involves complex processing - typically multiple stages of task-specific pre-training and fine-tuning. This has motivated researchers to explore simpler pipelines leveraging large language models (LLMs) that can work in a zero-shot manner. However, since zero-shot inference does not make use of a training set of pairs of queries and their relevant documents, its performance is mostly worse than that of supervised models, which are trained on such example pairs. Motivated by the existing findings that training examples generally improve zero-shot performance, in our work, we explore if this also applies to ranking models. More specifically, given a query and a pair of documents, the preference prediction task is improved by augmenting examples of preferences for similar queries from a training set. Our proposed pairwise few-shot ranker demonstrates consistent improvements over the zero-shot baseline on both in-domain (TREC DL) and out-domain (BEIR subset) retrieval benchmarks. Our method also achieves a close performance to that of a supervised model without requiring any complex training pipeline. © 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

FSM: A Fine-grained Splitting and Merging Framework for Dual-balanced Graph Partition 50th

FSM: A Fine-grained Splitting and Merging Framework for Dual...

引用

50th International Conference on Very Large data Bases, VLDB 2024

作者： Liu, Chengjun Peng, Zhuo Zheng, Weiguo Zou, Lei School of Data Science Fudan University China Wangxuan Institute of Computer Technology Peking University China

Partitioning a large graph into smaller subgraphs by minimizing the number of cutting vertices and edges, namely cut size or replication factor, plays a crucial role in distributed graph processing tasks. However, many prior works have primarily focused on optimizing the cut size by considering only vertex balance or edge balance, leading to significant workload imbalance and consequently hindering the performance of downstream tasks. Therefore, in this paper, we address the dual-balanced graph partition problem that minimizes the cut size while simultaneously guaranteeing both vertex and edge balance. We propose a lightweight effective twophase framework, namely fine-grained splitting and merging (FSM), which decomposes the graph into more and smaller partitions and then merges them. FSM offers the flexibility of integrating with various state-of-the-art single-balanced techniques. We develop two efficient algorithms Fast Merging and Precise Merging to enable trade-offs between computational efficiency and partitioning quality. Experimental results on large real-world graphs demonstrate that FSM achieves state-of-the-art cut size while maintaining dual balance. The runtime for downstream tasks PageRank, connected component, and diameter estimation, can be reduced by a large proportion, up to 9.43%, 11.35%, and 17.94%, respectively. © 2024, VLDB Endowment. All rights reserved.

关键词： Merging

来源：评论

学校读者我要写书评

暂无评论

Ontology-Based Decision-Making Support in Real-Time Strategy Game 4

Ontology-Based Decision-Making Support in Real-Time Strategy...

引用

4th International Conference on Consumer Electronics and Computer Engineering, ICCECE 2024

作者： Wang, Yaochu Yang, Jian Yang, Guoli School of Automation Science and Engineering South China University of Technology Guangzhou China Advanced Institute of Big Data Department of Big Data Intelligence Beijing China

ISBN: (纸本)9798350384062

Real-Time Strategy (RTS) games have attracted millions of players due to their characteristics of diverse scenes and flexible decision-making mechanisms. However, the mechanisms, contents, and operations of RTS games could be quite complex for beginners. The frustration and poor beginner experience caused by the high level of difficulty in getting started may lead novice players to give up on a specific RTS game due to a lack of sense of achievement. To address this issue and enhance the gaming experiences of beginners, this article proposes to assist beginners in familiarizing themselves with the mechanisms of the games and evaluating the opponents' abilities based on ontology. Firstly, an ontology will be built according to the properties of game units, which can provide beginners with a clear and comprehensive conceptual framework to understand game mechanisms and operations. Secondly, a knowledge-based reasoning method will be proposed to help beginners evaluate their opponents' abilities. Using a specific RTS game as a demonstration, the ontology is built and a reasoning example is proposed. By analyzing and reasoning about the abilities of their opponents' game units, this reasoning method can help beginners develop more effective strategies and tactics to enhance their experience in the game. © 2024 IEEE.

关键词： Ontology

来源：评论

学校读者我要写书评

暂无评论

Evaluation of neutron cross sections of^(48)Ti based on the Unified-Monte-Carlo-B method

引用

Chinese Physics C 2024年第7期48卷 118-130页

作者：夏候琼续瑞瑞刘萍 Roberto Capote 田源陶曦吴小飞张玥葛智刚舒能川冯松 China Nuclear Data Center China Institute of Atomic EnergyBeijing 102413China NAPC–Nuclear Data Section P.O.Box 100Vienna International CentreA-1400ViennaAustria School of Nuclear Science and Technology University of South ChinaHengyang 421001China

A cross section evaluation of neutron induced reactions on^(48)Ti is undertaken using the Unified Monte Carlo-B(UMC-B)*** evaluation concentrates on estimating the covariance and the use of the UMC-B allows avoiding the deficiencies of linear regression brought by the traditional least squares *** main neutron and charged particle emission reactions from n+^(48)Ti in the fast neutron energy region below 20 MeV are studied in this *** posterior probability density function(PDF)of each neutron cross section is obtained in a UMC-B Bayesian approach by convoluting the model PDFs sampled based on model parameters and the likelihood functions for the experimental *** model parameters including level density,pair corrections,optical model and Kalbach matrix element parameter are stochastically sampled with the assumption of normal distributions to estimate the model *** Cholesky factorization approach is applied to consider potential parameter ***,the posterior covariance matrices are generated using the UMC-B generated *** new evaluated results are compared with the CENDL-3.2,ENDF/B-VIII.0,JEFF-3.3,TENDL-2021 and JENDL-5 evaluations and differences are discussed.

关键词： covariance Unified Monte Carlo B(UMC-B) ^(48)Ti neutron cross sections Cholesky factor-ization approach

来源：评论

学校读者我要写书评

暂无评论

Bridging the Domain Gap in Grounded Situation Recognition via Unifying Event Extraction across Modalities

引用

data Intelligence 2025年第1期7卷 143-162页

作者： Qingwen Liu Zejun Li Zhihao Fan Cunxiang Yin Yancheng He Jing Cai Jinhua Fu Zhongyu Wei Institute of Science and Technology for Brain-Inspired Intelligence Fudan UniversityHandan RoadShanghai 200433China School of Data Science Fudan UniversityHandan RoadShanghai 200433China Tencent Nanshan StrictShenzhen 518057China

Event extraction extracts event frames from text, while grounded situation recognition detects events in images. As real-world applications frequently encounter a multitude of unforeseen events, certain researchers have introduced cross-domain and in-domain event extraction, while grounded situation recognition primarily explores in-domain scenarios. Therefore, in this paper, we propose cross-domain grounded situation recognition and establish a new benchmark SWiG-XD. In this more challenging setting, we deepen the connection between the two tasks based on their underlying unity in two different modalities and explore how to transfer the generalization ability from text to images. Firstly, we utilize ChatGPT to automatically generate textual data, which can be divided into two categories. One category is directly matched with images, establishing a direct connection with the images. The other category encompasses all event types and possesses greater generalization. Then we employ a unified model framework to establish the association between textual concepts and local image features and achieve cross-domain generalization transfer across modalities through modality-shared prompts and self-attention mechanism. Furthermore, we incorporate textual data with higher generalization to further assist in improving generalization on images. The experimental results on the newly constructed benchmark demonstrate the effectiveness of our method.

关键词： Event argument extraction Cross-domain generalization Unified cross-modal framework Modalityshared prompt Grounded situation recognition

来源：评论

学校读者我要写书评

暂无评论

Knowledge Leaks in data-Driven Business Models? Exploring Different Types of Knowledge Risks and Protection Measures

引用

Schmalenbach Journal of Business Research 2024年第3期76卷 357-396页

作者： Fruhwirth, Michael Pammer-Schindler, Viktoria Thalmann, Stefan Know-Center GmbH Institute for Interactive Systems and Data Science Faculty of Computer Science and Biomedical Engineering Graz University of Technology Silicon Austria Labs GmbH Business Analytics and Data Science-Center (BANDAS-Center) School of Business Economics and Social Sciences University of Graz

data-driven business models imply the inter-organisational exchange of data or similar value objects. data science methods enable organisations to discover patterns and eventually knowledge from data. Further, by training machine learning models, knowledge is materialised in those models. Thus, organisations might risk the exposure of competitive knowledge by sharing data-related value objects, such as data, models or predictions. Although knowledge risks have been studied in traditional business models, little research has been conducted in the direction of data-driven business models. In this explorative qualitative study, we conducted 28 expert interviews in three rounds (two exploratory and one evaluatory) and identified five types of risks along the three basic types of value objects: data, models and predictions. These risks depend on the context, i.e., when competitive knowledge could be discovered from shared value objects. We found that those risks can be mitigated by technology, contractual regulations, trusted relationships, and adjusting the business model design. In this study, we show that the risk of knowledge leakage is a relevant risk factor in data-driven business models. Overall, knowledge risks should be considered already during business model design, and their management requires an interdisciplinary approach via a balanced assessment. The level of knowledge protection from a technology perspective highly depends on computer science innovations and thus is a moving target. As an outlook, we suggest that knowledge risk will become even more relevant with the extensive usage of machine learning and artificial intelligence in data-driven business models. © The Author(s) 2024.

关键词： Business model innovation data analytics data-driven business models Knowledge risks Risk management Value objects

来源：评论

学校读者我要写书评

暂无评论

AW-YOLOv9: Adverse Weather Conditions Adaptation for UAV Detection 24th

AW-YOLOv9: Adverse Weather Conditions Adaptation for UAV De...

引用

24th International Conference on Algorithms and Architectures for Parallel Processing, ICA3PP 2024

作者： Yang, Runjie Guo, Yongde Yan, Jun Yang, Haonan Dai, Zifeng Faculty of Data Science City University of Macau SAR Macau China School of Data Science Qingdao University of Science and Technology Qingdao266000 China Zhuhai Aerospace Microchips Science and Technology Co. Ltd. Zhuhai519000 China School of Cyberspace Science and Technology Beijing Institute of Technology Beijing100039 China

ISBN: (纸本)9789819615445

Effective and accurate detection of unmanned aerial vehicles (UAVs) is crucial for combating malicious UAV systems. However, adverse weather conditions, such as haze or low light, often degrade the quality of captured UAV images. Traditional target detection models applied to these low-quality UAV images usually suffer from decreased accuracy. To address this challenge, this paper proposes AW-YOLOv9, a UAV target detection method designed for adverse weather conditions. Specifically, an image enhancement processing (IEP) module with adaptive learning parameters is introduced to suppress weather-related interference in the images and restore hidden target information. The DP module embedded within YOLOv9 was designed to reduce the network’s parameter count while improving the detection speed. We established a dataset of UAV images under adverse weather conditions and utilized this dataset to validate the superiority of AW-YOLOv9. The experiments demonstrate that the proposed method achieves a 3.30% improvement in mean average precision compared to SOTA methods, effectively enhancing the detection accuracy of UAV targets under adverse weather conditions. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Aircraft detection

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：