检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Zhang, Zhihao Zhao, Jun Zhang, Qi Gui, Tao Huang, Xuanjing School of Computer Science Fudan University China Institute of Modern Languages and Linguistics Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China

Large Language Models (LLMs) have demonstrated considerable cross-lingual alignment and generalization ability. Current research primarily focuses on improving LLMs' crosslingual generalization capabilities. However, there is still a lack of research on the intrinsic mechanisms of how LLMs achieve crosslingual alignment. From the perspective of region partitioning, this paper conducts several investigations on the linguistic competence of LLMs. We discover a core region in LLMs that corresponds to linguistic competence, accounting for approximately 1% of the total model parameters. Removing this core region by setting parameters to zero results in a significant performance decrease across 30 different languages. Furthermore, this core region exhibits significant dimensional dependence, perturbations to even a single parameter on specific dimensions leading to a loss of linguistic competence. Moreover, we discover that distinct monolingual regions exist for different languages, and disruption to these specific regions substantially reduces the LLMs' proficiency in those corresponding languages. Our research also indicates that freezing the core linguistic region during further pre-training can mitigate the issue of catastrophic forgetting (CF), a common phenomenon observed during further pre-training of LLMs. Overall, exploring the LLMs' functional regions provides insights into the foundation of their intelligence. Copyright © 2024, The Authors. All rights reserved.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

Advancements and Challenges in Space Exploration and Cyber Threat Mitigation Techniques: A Comprehensive Review 57

Advancements and Challenges in Space Exploration and Cyber T...

引用

57th IAA Symposium on Safety, Quality and Knowledge Management in Space Activities at the 75th International Astronautical Congress, IAC 2024

作者： Almousa, May M. Abd El-Latif, Ahmed A. Information Technology Department College of Computer and Information Sciences Princess Nourah Bint Abdulrahman University Riyadh11671 Saudi Arabia EIAS Data Science Lab College of Computer and Information Sciences Center of Excellence in Quantum and Intelligent Computing Prince Sultan University Riyadh11586 Saudi Arabia

ISBN: (纸本)9798331312275

Space exploration has been revolutionized by deep learning (DL) and machine learning (ML) technologies, enabling advanced data analysis, spacecraft navigation, and mission planning. However, the increasing reliance on digital systems also introduces new cybersecurity vulnerabilities. This research presents a comprehensive literature review of how ML and DL can enhance mission effectiveness while mitigating these risks. By examining over 50 research articles, we identify significant advancements and challenges in space mission cybersecurity, with a particular focus on DL for threat detection, risk assessment, and system protection. Our findings highlight the potential of DL to address emerging cybersecurity threats and improve the overall security and efficiency of space missions. This work aims to provide a foundation for developing robust and secure AI-powered systems for future space exploration endeavors. Copyright © 2024 by the International Astronautical Federation (IAF).

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

Objective evaluation of relighting models on translucent materials from multispectral RTI images

Objective evaluation of relighting models on translucent mat...

引用

2021 Material Appearance Conference, MAAP 2021

作者： Kitanovski, Vlado Hardeberg, Jon Yngve The Norwegian Color and Visual Computing Laboratory Department of Computer Science NTNU - Norwegian University of Science and Technology Gjøvik Norway

In this paper, we evaluate the quality of reconstruction i.e. relighting from images obtained by a newly developed multispectral reflectance transformation imaging (MS-RTI) system. The captured MS-RTI images are of objects with different translucency and color. We use the most common methods for relighting the objects: polynomial texture mapping (PTM) and hemispherical harmonics (HSH), as well as the recent discrete model decomposition (DMD). The results show that all three models can reconstruct the images of translucent materials, with the reconstruction error varying with translucency but still in the range of what has been reported for other non-translucent materials. DMD relighted images are marginally better for the most transparent objects, while HSH- and PTM- relighted images appear to be better for the opaquer objects. The estimation of the surface normals of highly translucent objects using photometric stereo is not very accurate. Utilizing the peak of the fitted angular reflectance field, the relighting models, especially PTM, can provide more accurate estimation of the surface normals. © 2021 Society for Imaging science and Technology. All rights reserved.

关键词： Reflection

来源：评论

学校读者我要写书评

暂无评论

Hashgraph consensus-based simultaneous, asynchronous and big-data supported blockchain mechanism for IT-OT enabled critical infrastructure sectors

引用

Journal of Ambient Intelligence and Humanized computing 2025年 1-13页

作者： Hasan, Mohammad Kamrul Ghazal, Taher M. Issa, Ghassan Islam, Shayla Ismail, Ahmad Fadzil Kabir, S. Rayhan Bangi43600 Malaysia Hourani Center for Applied Scientific Research Al-Ahliyya Amman University Amman11937 Jordan School of Computing Skyline University College University City Sharjah Skyline University College Sharjah1797 United Arab Emirates Institue of Computer Science and Digitial Innovation UCSI University Malaysia Kuala Lumpur56000 Malaysia Department of Electrical and Computer Engineering International Islamic University of Malaysia Kuala Lumpur53000 Malaysia

Critical Infrastructure (CI) refers to the essential areas made up of public, private, and business sectors for the security of a country's development, such as electricity, water, health, education, etc. where information technologies (ITs) and operational technologies (OTs) enable automation services of CI. Currently, blockchain is being experimented in various CI sectors and any CI sector accumulates the data of millions of people in a country or nation from which big data is being generated. Big data connections are seen with large-scale blockchain networks because as the size of the blockchain network increases, so does the amount of connected data. Typically, the blockchain does not support asynchronous and simultaneous processes that indicate multiple block compilation does not allow at the same time for the blockchain systems. That is why high energy consumption, high bandwidth dependency, long-time computation, high mining process, and system slowness issues are seen on large-scale blockchain networks where big data is also connected. However, Hashgraph is an alternative distributed ledger technology (DLT) to blockchain and is also considered a consensus method that supports Asynchronous Byzantine Fault Tolerance (ABFT) as well as simultaneous process. Moreover, Hashgraph does not need mining and network node verification processes which saves bandwidth. In this research, we proposed a Hashgraph-enabled blockchain model for reducing the above issues and illustrated the adoption of IT-enabled OT to support CI. In addition, this manuscript discusses how the proposed model can create positive impacts in CI areas by replacing traditional blockchain systems. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Critical infrastructures

来源：评论

学校读者我要写书评

暂无评论

Analytical Study of the Features and Parameters of Off-Angle Iris Image vs. Deep Learning Model

Analytical Study of the Features and Parameters of Off-Angle...

引用

2024 International Conference on Decision Aid sciences and Applications, DASA 2024

作者： Suleiman, Aghiad Khedr Shafek, Darin Mahfuri, Mahmoud Khishe, Mohammad Saeed, Ali Ghwanmeh, Sameh Alma'Moon University College Cybersecurity and Cloud Computing Technologies Engineering Dept. Baghdad Iraq Alma'Moon University College Computer Engineering Techniques Dept. Baghdad Iraq International University for Science and Technology Computer and Information Technology Department Syria Innovation Center for Artificial Intelligence Applications Yuan Ze University Taiwan Jadara Research Center Jadara University Jordan Jordan Computer Center Northern Technical University Ninevah Iraq College of Computer Information Technology Computer Science Dept American University in the Emirates United Arab Emirates

ISBN: (纸本)9798350369106

This paper presents an analytical study and a technique for extracting the features of a common case of images of the iris called off-angle iris which was taken for persons identification systems. The main problem when using biological iris measurements to identify the persons is the difficulty of identifying and extracting features of the iris. This problem is increasing when dealing with off-angle iris and it leads to decreased system accuracy and increased system rate error. In return, all the transfer learning techniques face difficulties in the case of heavily degraded data and partially occluded, and off-angle. A new method has a whole new methodology to deal with the image as it is without transformation processes. A new algorithm has been included for extracting features of the iris through the pupil switching points. The most discriminating points of the iris depend on the biological human eye statistics and analytical study. It has been trained and tested on the common images of the off-angle iris database so-called: 'CASIA Iris 1.0'. It has been implemented in the MATLAB environment. The results showed the efficiency of this technique, high precision and more importantly low failed acceptance rate. It emphasized that it is adaptive as well as efficiency improvement of the system. © 2024 IEEE.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

ToolSword: Unveiling Safety Issues of Large Language Models in Tool Learning Across Three Stages

arXiv

引用

arXiv 2024年

作者： Ye, Junjie Li, Sixian Li, Guanyu Huang, Caishuang Gao, Songyang Wu, Yilong Zhang, Qi Gui, Tao Huang, Xuanjing School of Computer Science Fudan University China Institute of Modern Languages and Linguistics Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China

Tool learning is widely acknowledged as a foundational approach for deploying large language models (LLMs) in real-world scenarios. While current research primarily emphasizes leveraging tools to augment LLMs, it frequently neglects emerging safety considerations tied to their application. To fill this gap, we present ToolSword, a comprehensive framework dedicated to meticulously investigating safety issues linked to LLMs in tool learning. Specifically, ToolSword delineates six safety scenarios for LLMs in tool learning, encompassing malicious queries and jailbreak attacks in the input stage, noisy misdirection and risky cues in the execution stage, and harmful feedback and error conflicts in the output stage. Experiments conducted on 11 open-source and closed-source LLMs reveal enduring safety challenges in tool learning, such as handling harmful queries, employing risky tools, and delivering detrimental feedback, which even GPT-4 is susceptible to. Moreover, we conduct further studies with the aim of fostering research on tool learning safety. The data is released in https://***/Junjie-Ye/ToolSword. Copyright © 2024, The Authors. All rights reserved.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Extracting Information from Drug Label Using Image and Text Processing

Extracting Information from Drug Label Using Image and Text ...

引用

International Joint Conference on computer science and Software Engineering (JCSSE)

作者： Peeraporn Rimdusit Phakhatorn Aphiromyarnont Suphakant Phimoltares Sasipa Panthuwadeethorn Department of Mathematics and Computer Science Faculty of Science Advanced Virtual and Intelligent Computing (AVIC) Center Chulalongkorn University Bangkok Thailand

ISBN: (数字)9798350381764

ISBN: (纸本)9798350381771

The purpose of this research aims to develop a method utilizing image and text processing to extract information from drug labels. The input images used in this research are drug labels on the packaging in the form of boxes, blister packs, sachets, or bottles of oral drugs that can be purchased at general drugstores without a prescription, in JPEG, JPG, or PNG format. The extracted information needs to be displayed on the drug label and visible in the image. The process involves image processing, optical character recognition, text processing, and text-to-speech. Experimental results show that the proposed method is effective in extracting information from drug labels with 80% accuracy in drug information searching from drug data file, fair and excellent ranges in information scoring, and 55.6% accuracy of the expiration date retrieval. The best results are obtained from sharp images of drug labels on boxes with standard fonts and clear details. The proposed method can be further developed into an application in the future.

关键词： Drugs Accuracy Image processing Optical character recognition Transform coding Reflection Text to speech

来源：评论

学校读者我要写书评

暂无评论

IoT-Cloud Assisted Botnet Detection Using Rat Swarm Optimizer with Deep Learning

引用

computers, Materials & Continua 2023年第2期74卷 3085-3100页

作者： Saeed Masoud Alshahrani Fatma S.Alrayes Hamed Alqahtani Jaber S.Alzahrani Mohammed Maray Sana Alazwari Mohamed A.Shamseldin Mesfer Al Duhayyim Department of Computer Science College of Computing and Information TechnologyShaqra UniversityShaqraSaudi Arabia Department of Information Systems College of Computer and Information SciencesPrincess Nourah Bint Abdulrahman UniversityP.O.Box 84428Riyadh11671Saudi Arabia Department of Information Systems College of Computer ScienceCenter of Artificial IntelligenceUnit of CybersecurityKing Khalid UniversityAbhaSaudi Arabia Department of Industrial Engineering College of Engineering at AlqunfudahUmm Al-Qura UniversitySaudi Arabia Department of Information Systems College of Computer ScienceKing Khalid UniversityAbhaSaudi Arabia Department of Information Technology College of Computers and Information TechnologyTaif UniversityTaif P.O.Box 11099Taif21944Saudi Arabia Department of Mechanical Engineering Faculty of Engineering and TechnologyFuture University in EgyptNew Cairo11835Egypt Department of Computer Science College of Sciences and Humanities-AflajPrince Sattam bin Abdulaziz UniversitySaudi Arabia

Nowadays,Internet of Things(IoT)has penetrated all facets of human life while on the other hand,IoT devices are heavily prone to *** has become important to develop an accurate system that can detect malicious attacks on IoT environments in order to mitigate security *** is one of the dreadfulmalicious entities that has affected many users for the past few *** is challenging to recognize Botnet since it has excellent carrying and hidden *** approaches have been employed to identify the source of Botnet at earlier *** Learning(ML)and Deep Learning(DL)techniques are developed based on heavy influence from Botnet detection *** spite of this,it is still a challenging task to detect Botnet at early stages due to low number of features accessible from Botnet *** current study devises IoT with Cloud Assisted Botnet Detection and Classification utilizingRat SwarmOptimizer with Deep Learning(BDC-RSODL)*** presented BDC-RSODL model includes a series of processes like pre-processing,feature subset selection,classification,and parameter ***,the network data is pre-processed to make it compatible for further ***,RSO algorithm is exploited for effective selection of subset of ***,Long Short TermMemory(LSTM)algorithm is utilized for both identification and classification of ***,Sine Cosine Algorithm(SCA)is executed for fine-tuning the hyperparameters related to LSTM *** order to validate the promising 3086 CMC,2023,vol.74,no.2 performance of BDC-RSODL system,a comprehensive comparison analysis was *** obtained results confirmed the supremacy of BDCRSODL model over recent approaches.

关键词： Internet of things cloud computing long short termmemory deep learning sine cosine algorithm feature selection

来源：评论

学校读者我要写书评

暂无评论

MRN: Multiplexed Routing Network for Incremental Multilingual Text Recognition

arXiv

引用

arXiv 2023年

作者： Zheng, Tianlun Chen, Zhineng Huang, Bingchen Zhang, Wei Jiang, Yu-Gang School of Computer Science Fudan University China Shanghai Collaborative Innovation Center of Intelligent Visual Computing China Gaoding AI China

Multilingual text recognition (MLTR) systems typically focus on a fixed set of languages, which makes it difficult to handle newly added languages or adapt to ever-changing data distribution. In this paper, we propose the Incremental MLTR (IMLTR) task in the context of incremental learning (IL), where different languages are introduced in batches. IMLTR is particularly challenging due to rehearsal-imbalance, which refers to the uneven distribution of sample characters in the rehearsal set, used to retain a small amount of old data as past memories. To address this issue, we propose a Multiplexed Routing Network (MRN). MRN trains a recognizer for each language that is currently seen. Subsequently, a language domain predictor is learned based on the rehearsal set to weigh the recognizers. Since the recognizers are derived from the original data, MRN effectively reduces the reliance on older data and better fights against catastrophic forgetting, the core issue in IL. We extensively evaluate MRN on MLT17 and MLT19 datasets. It outperforms existing general-purpose IL methods by large margins, with average accuracy improvements ranging from 10.3% to 35.8% under different settings. Code is available at https://***/simplify23/MRN. Copyright © 2023, The Authors. All rights reserved.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

Beyond Discrete Selection: Continuous Embedding Space Optimization for Generative Feature Selection 23

Beyond Discrete Selection: Continuous Embedding Space Optimi...

引用

23rd IEEE International Conference on Data Mining, ICDM 2023

作者： Xiao, Meng Wang, Dongjie Wu, Min Wang, Pengfei Zhou, Yuanchun Fu, Yanjie Chinese Academy of Sciences Computer Network Information Center China University of Chinese Academy of Sciences China Singapore University of Central Florida Department of Computer Science United States Arizona State University School of Computing and Ai United States

ISBN: (纸本)9798350307887

The goal of Feature Selection - comprising filter, wrapper, and embedded approaches - is to find the optimal feature subset for designated downstream tasks. Nevertheless, current feature selection methods are limited by: 1) the selection criteria of these methods are varied for different domains, leading them hard to be generalized;2) the selection performance of these approaches drops significantly when processing high-dimensional feature space coupled with small sample size. In light of these challenges, we pose the question: can selected feature subsets be more robust, accurate, and input dimensionality agnostic? In this paper, we reformulate the feature selection problem as a deep differentiable optimization task and propose a new research perspective: conceptualizing discrete feature subsetting as continuous embedding space optimization. We introduce a novel and principled framework that encompasses a sequential encoder, an accuracy evaluator, a sequential decoder, and a gradient ascent optimizer. This comprehensive framework includes four important steps: preparation of features-accuracy training data, deep feature subset embedding, gradient-optimized search, and feature subset reconstruction. Specifically, we utilize reinforcement feature selection learning to generate diverse and high-quality training data and enhance generalization. By optimizing reconstruction and accuracy losses, we embed feature selection knowledge into a continuous space using an encodere-valuator-decoder model structure. We employ a gradient ascent search algorithm to find better embeddings in the learned embedding space. Furthermore, we reconstruct feature selection solutions using these embeddings and select the feature subset with the highest performance for downstream tasks as the optimal subset. Finally, extensive experimental results demonstrate the effectiveness of our proposed method, showcasing significant enhancements in feature selection robustness and accuracy. To improve t

关键词： Feature Selection

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：