检索结果-内蒙古大学图书馆

50th Scientific Meeting of the Italian Statistical Society, SIS 2021

作者： Del Sarto, Simone Coppola, Paolo Troìa, Matteo Department of Political Science University of Perugia Perugia Italy Department of Mathematics Computer Science and Physics University of Udine Udine Italy Capgemini Italia – Insights & Data Milan Italy

ISBN: (纸本)9783031166082

Public procurement is a field particularly prone to corrupt acts. In this regard, red flag indicators are developed with the purpose of signalling anomalies and alerting the system for a potential risk of corruption. By exploiting data coming from the Italian National database of Public Contracts, a set of red flag indicators for the measurement of the risk of corruption in public procurement are computed for a sample of contracting authorities and monitored over time. Afterwards, a latent Markov model for continuous responses is applied to the data at issue, in order to: i. identify groups of contracting administrations (i.e., the model latent states), characterised by different behaviours on the basis of the selected red flags, and ii. estimate the transitions over time across the ascertained groups. Results show that seven profiles of administrations may be highlighted on account of the selected red flags. Among them, four profiles include contracting authorities with indicator values associable to a low risk, whereas one profile can be labelled as the most at-risk one according to the respective indicator means. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Crime

来源：评论

学校读者我要写书评

暂无评论

ST-GNN-Powered Compression for Latency Reduction in Hybrid Satellite-Terrestrial Networks 8

ST-GNN-Powered Compression for Latency Reduction in Hybrid S...

引用

8th International Conference on Information, Control, and Communication Technologies, ICCT 2024

作者： Do, Phuc Hao Due Le, Tran Berezkin, Aleksandr Kirichek, Ruslan The Bonch-Bruevich Saint - Petersburg State University of Telecommunications Faculty of Infocommunication Networks and Systems Saint-Petersburg Russia University of Wisconsin-Stout Department of Mathematics Statistics & Computer Science MenomonieWI United States

ISBN: (纸本)9798331517564

Hybrid Satellite-Terrestrial Communication Networks (HSTCNs) have emerged as a critical infrastructure for modern communication, integrating the capabilities of satellite and terrestrial networks to meet the increasing demands for connectivity and reliability. However, enhancing data transmission efficiency in such heterogeneous networks remains challenging, particularly concerning latency. This research aims to address this challenge by proposing a lossless compression method leveraging custom Spatio-Temporal Graph Neural Network (ST-GNN) architectures to reduce data transmission latency in HSTCNs. The primary objectives of this study are twofold: first, to evaluate the compression rates achieved by different ST-GNN-based network traffic compression methods, and second, to conduct a comparative analysis of data transmission latency over HSTCNs based on the compression of traffic networks. By exploiting the inherent spatio-temporal characteristics of communication data, ST- GNNs offer a promising approach to compress network traffic while minimizing latency effectively. The empirical assessment demonstrates the effectiveness of the proposed ST-GNN-based compression methods in achieving competitive compression rates compared to existing techniques. Furthermore, the comparative analysis highlights the potential of ST-GNNs in improving data transmission efficiency and reducing latency in HSTCNs. © 2024 IEEE.

关键词： Satellite communication systems

来源：评论

学校读者我要写书评

暂无评论

Early Determination for Intra Block Copy Prediction with Refined Screen Content Detection

Early Determination for Intra Block Copy Prediction with Ref...

引用

IEEE Visual Communications and Image Processing (VCIP)

作者： Zhenhao Sun Meng Wang Yingwen Zhang Shiqi Wang Sam Kwong Department of Computer Science City University of Hong Kong China School of Data Science Lingnan University China

ISBN: (数字)9798331529543

ISBN: (纸本)9798331529550

The intra block copy (IBC) mode introduced can significantly improve the coding efficiency for screen content. However, with the mixed content, the coding performance improvement of IBC is limited and while also introducing the increases of the coding complexity. This paper proposes an early determination scheme for IBC prediction to reduce the encoding complexity over mixed content sequences containing captured content regions. Our method is based on a refined CTU-level screen content detection method, which can better distinguish screen content in mixed scenes. Experimental results show that the proposed method reduces encoding complexity with moderate coding performance loss on mixed content sequences.

关键词： Video coding Image coding Visual communication Semantics Text detection Complexity theory

来源：评论

学校读者我要写书评

暂无评论

Comparison of Decision Tree and Support Vector Machine for Predicting Jakarta Air Quality Index

Comparison of Decision Tree and Support Vector Machine for P...

引用

computer science, Information Technology and Engineering (ICCoSITE), International Conference on

作者： Musyaffa Ayman Rafif Gede Sanjaya Indrajaya Muhammad Kent Al-Ghazi Johnny Johnny Noviyanti T M Sagala Data Science Program School of Computer Sciences Bina Nusantara University Jakarta Indonesia Statistics Department School of Computer Science Bina Nusantara University Jakarta Indonesia

The people in the capital city of the Republic of Indonesia, Jakarta, have been living for years with air pollution. Air quality has been a concern for a long time due to the health risks it has on people, especially those at risk. Using classification algorithms, we would like to implement data mining for the air quality index and make predictions from our capital city, DKI Jakarta. This research aims to compare Decision Tree and Support Vector Machine in predicting Air Quality Index in DKI Jakarta in 2020. The dataset was collected from the Jakarta Open data Web, and the results of this research highlight the effectiveness of data mining techniques in predicting air quality index levels in Jakarta and emphasize the importance of selecting the appropriate classification algorithm for data mining applications. This research shows that SVM, compared to the Decision tree, has a better result with an accuracy score of 87.86%. On the other hand, Decision Tree has a score of 90.56%.

关键词： Support vector machines Urban areas Support vector machine classification Predictive models Air pollution Classification algorithms Decision trees

来源：评论

学校读者我要写书评

暂无评论

An Open-source UAV Digital Twin framework: A Case Study on Remote Sensing in the Andean Mountains

引用

Journal of Intelligent and Robotic Systems: Theory and Applications 2025年第2期111卷 1-19页

作者： Valencia, Esteban Toapanta, Francisco Oña, Gabriel Carrillo, Andrey Aláez, Daniel Loyaga, Erick Chamorro, William Cruz, Patricio Abad, Jackeline Vandewalle, Patrick Laboratory of UAS ATA Research group Mechanical Engineering Department Escuela Politécnica Nacional Quito Ecuador Department of Statistics Computer Science and Mathematics Universidad Pública de Navarra Pamplona31006 Spain Institute of Smart Cities Universidad Pública de Navarra Pamplona31006 Spain KU Leuven Jan Pieter de Nayerlaan 5 Sint-Katelijne-Waver Belgium

The increasing demand for unmanned aerial vehicles (UAVs) in the aerospace industry highlights the need for precise simulation environments, especially in remote regions. This study develops an open-source framework for a customized UAV simulation environment using ROS-Gazebo and Ardupilot. The method includes a realistic reconstruction of the environment based on free satellite information, the ROS-Gazebo scheme for modeling and testing the UAV platform, and Ardupilot for mission design and deployment. The methodology aims to reduce costs in surveillance missions and conduct more efficient operations in high-risk and challenging areas. To evaluate the accuracy of UAV positioning for simulating real missions, an experimental validation of the digital twin was carried out using real flight data records from two self-built UAVs equipped with the open-source Ardupilot autopilot. One of the aircraft, a fixed-wing UAV, was tested near the Antisana volcano (4500 meters above sea level), where wind gusts reached speeds of 9 to 12 m/s. These tests revealed maximum errors of 9% in the Z-axis trajectory (altitude). The other aircraft, a quadcopter, was evaluated in the Parque Carolina in Quito (2800 meters above sea level), with wind gusts between 5 and 7 m/s, showing an error of 8% in the Z-axis trajectory. The presented results demonstrate the suitability of the proposed method for emulating complex missions using digital twin models. In this regard, the main contribution of this work lies in its potential for the precise prediction of flight missions in the Z-axis, a crucial variable for avoiding collisions in mountainous regions. © The Author(s) 2025.

关键词： Aerospace industry

来源：评论

学校读者我要写书评

暂无评论

TRACE: TRansformer-based Attribution using Contrastive Embeddings in LLMs

arXiv

引用

arXiv 2024年

作者： Wang, Cheng Lu, Xinyang Ng, See-Kiong Low, Bryan Kian Hsiang Department of Computer Science National University of Singapore Singapore Institute of Data Science National University of Singapore Singapore

The rapid evolution of large language models (LLMs) represents a substantial leap forward in natural language understanding and generation. However, alongside these advancements come significant challenges related to the accountability and transparency of LLM responses. Reliable source attribution is essential to adhering to stringent legal and regulatory standards, including those set forth by the General data Protection Regulation. Despite the well-established methods in source attribution within the computer vision domain, the application of robust attribution frameworks to natural language processing remains underexplored. To bridge this gap, we propose a novel and versatile TRansformer-based Attribution framework using Contrastive Embeddings called TRACE that, in particular, exploits contrastive learning for source attribution. We perform an extensive empirical evaluation to demonstrate the performance and efficiency of TRACE in various settings and show that TRACE significantly improves the ability to attribute sources accurately, making it a valuable tool for enhancing the reliability and trustworthiness of LLMs. © 2024, CC BY.

关键词： Embeddings

来源：评论

学校读者我要写书评

暂无评论

Continuous Partitioning for Graph-Based Semi-Supervised Learning 38

Continuous Partitioning for Graph-Based Semi-Supervised Lear...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Holtz, Chester Chen, Pengwen Wan, Zhengchao Cheng, Chung-Kuan Mishne, Gal Halicioǧlu Data Science Institute University of California La Jolla San DiegoCA United States Department of Applied Mathematics National Chung Hsing University South District Taichung Taiwan Department of Mathematics University of Missouri ColumbiaMO United States Department of Computer Science University of California La Jolla San DiegoCA United States

Laplace learning algorithms for graph-based semi-supervised learning have been shown to suffer from degeneracy at low label rates and in imbalanced class regimes. Here, we propose CutSSL: a framework for graph-based semi-supervised learning based on continuous nonconvex quadratic programming, which provably obtains integer solutions. Our framework is naturally motivated by an exact quadratic relaxation of a cardinality-constrained minimum-cut graph partitioning problem. Furthermore, we show our formulation is related to an optimization problem whose approximate solution is the mean-shifted Laplace learning heuristic, thus providing new insight into the performance of this heuristic. We demonstrate that CutSSL significantly surpasses the current state-of-the-art on k-nearest neighbor graphs and large real-world graph benchmarks across a variety of label rates, class imbalance, and label imbalance regimes. Our implementation is available on github. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Generalization of the Persistent Laplacian to Simplicial Maps 39

A Generalization of the Persistent Laplacian to Simplicial M...

引用

39th International Symposium on Computational Geometry, SoCG 2023

作者： Gülen, Aziz Burak Mémoli, Facundo Wan, Zhengchao Wang, Yusu Department of Mathematics The Ohio State University ColumbusOH United States Department of Mathematics Department of Computer Science and Engineering The Ohio State University ColumbusOH United States Halıcıoğlu Data Science Institute University of California San DiegoCA United States

ISBN: (纸本)9783959772730

The (combinatorial) graph Laplacian is a fundamental object in the analysis of, and optimization on, graphs. Via a topological view, this operator can be extended to a simplicial complex K and therefore offers a way to perform "signal processing" on p-(co)chains of K. Recently, the concept of persistent Laplacian was proposed and studied for a pair of simplicial complexes K, → L connected by an inclusion relation, further broadening the use of Laplace-based operators. In this paper, we significantly expand the scope of the persistent Laplacian by generalizing it to a pair of weighted simplicial complexes connected by a weight preserving simplicial map f : K → L. Such a simplicial map setting arises frequently, e.g., when relating a coarsened simplicial representation with an original representation, or the case when the two simplicial complexes are spanned by different point sets, i.e. cases in which it does not hold that K ⊂ L. However, the simplicial map setting is much more challenging than the inclusion setting since the underlying algebraic structure is much more complicated. We present a natural generalization of the persistent Laplacian to the simplicial setting. To shed insight on the structure behind it, as well as to develop an algorithm to compute it, we exploit the relationship between the persistent Laplacian and the Schur complement of a matrix. A critical step is to view the Schur complement as a functorial way of restricting a self-adjoint positive semi-definite operator to a given subspace. As a consequence of this relation, we prove that the qth persistent Betti number of the simplicial map f : K → L equals the nullity of the qth persistent Laplacian ∆K, Lq. We then propose an algorithm for finding the matrix representation of ∆K, Lq which in turn yields a fundamentally different algorithm for computing the qth persistent Betti number of a simplicial map. Finally, we study the persistent Laplacian on simplicial towers under weight-preserving simpli

关键词： Matrix algebra

来源：评论

学校读者我要写书评

暂无评论

Do You Trust Your Model? Emerging Malware Threats in the Deep Learning Ecosystem

arXiv

引用

arXiv 2024年

作者： Hitaj, Dorjan Pagnotta, Giulio De Gaspari, Fabio Ruko, Sediola Hitaj, Briland Mancini, Luigi V. Perez-Cruz, Fernando Department of Computer Science Sapienza University of Rome Italy Computer Science Laboratory SRI International Swiss Data Science Center and Computer Science Department ETH Zürich DEIM Università degli Studi della Tuscia Italy

Training high-quality deep learning models is a challenging task due to computational and technical requirements. A growing number of individuals, institutions, and companies increasingly rely on pre-trained, third-party models made available in public repositories. These models are often used directly or integrated in product pipelines with no particular precautions, since they are effectively just data in tensor form and considered safe. In this paper, we raise awareness of a new machine learning supply chain threat targeting neural networks. We introduce MaleficNet 2.0, a novel technique to embed self-extracting, self-executing malware in neural networks. MaleficNet 2.0 uses spread-spectrum channel coding combined with error correction techniques to inject malicious payloads in the parameters of deep neural networks. MaleficNet 2.0 injection technique is stealthy, does not degrade the performance of the model, and is robust against removal techniques. We design our approach to work both in traditional and distributed learning settings such as Federated Learning, and demonstrate that it is effective even when a reduced number of bits is used for the model parameters. Finally, we implement a proof-of-concept self-extracting neural network malware using MaleficNet 2.0, demonstrating the practicality of the attack against a widely adopted machine learning framework. Our aim with this work is to raise awareness against these new, dangerous attacks both in the research community and industry, and we hope to encourage further research in mitigation techniques against such threats. © 2024, CC BY-NC-SA.

关键词： Malware

来源：评论

学校读者我要写书评

暂无评论

Residual Learning Inspired Crossover Operator and Strategy Enhancements for Evolutionary Multitasking

arXiv

引用

arXiv 2025年

作者： Wang, Ruilin Feng, Xiang Yu, Huiqun Lai, Edmund M.-K. Department of Computer Science and Engineering East China University of Science and Technology Shanghai China Department of Data Science and Artificial Intelligence Auckland University of Technology Auckland New Zealand

In evolutionary multitasking, strategies such as crossover operators and skill factor assignment are critical for effective knowledge transfer. Existing improvements to crossover operators primarily focus on low-dimensional variable combinations, such as arithmetic crossover or partially mapped crossover, which are insufficient for modeling complex high-dimensional ***, static or semi-dynamic crossover strategies fail to adapt to the dynamic dependencies among tasks. In addition, current Multifactorial Evolutionary Algorithm frameworks often rely on fixed skill factor assignment strategies, lacking flexibility. To address these limitations, this paper proposes the Multifactorial Evolutionary Algorithm-Residual Learning (MFEA-RL) method based on residual learning. The method employs a Very Deep Super-Resolution (VDSR) model to generate high-dimensional residual representations of individuals, enhancing the modeling of complex relationships within dimensions. A ResNet-based mechanism dynamically assigns skill factors to improve task adaptability, while a random mapping mechanism efficiently performs crossover operations and mitigates the risk of negative transfer. Theoretical analysis and experimental results show that MFEA-RL outperforms state-of-the-art multitasking algorithms. It excels in both convergence and adaptability on standard evolutionary multitasking benchmarks, including CEC2017-MTSO and WCCI2020-MTSO. Additionally, its effectiveness is validated through a real-world application scenario. Copyright © 2025, The Authors. All rights reserved.

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：