检索结果-内蒙古大学图书馆

2024 International Symposium on Internet of Things and Smart Cities, ISITSC 2024

作者： Liu, Zhao Liu, Wei Chen, Junyang School of Artificial Intelligence Sun Yat-sen University Zhuhai China College of Computer Science and Software Engineering Shenzhen University Shenzhen China

ISBN: (纸本)9798331505516

Predicting the next Point-of-Interest (POI) is crucial for location-based services. In this paper, we propose the Time-enhanced Sequence Prediction Model (TSPM) to improve the accuracy of next POI recommendations by incorporating temporal information and dynamic graph *** approach utilizes a Time-enhanced Sequence-based Dynamic Graph (TSDG) that captures both temporal transitions of POIs and sequential dependencies in user behavior. By embedding temporal information directly into the graph structure, TSPM effectively models user movements. We further enhance POI embeddings using knowledge graph techniques and Eigenmap to preserve the topological properties of the *** proposed model integrates these enriched embeddings into a Time-aware Recurrent Neural Network (TiRNN) to capture the influence of past check-ins across different time intervals. Experiments on real-world datasets demonstrate that TSPM significantly outperforms existing methods in prediction accuracy. © 2024 IEEE.

关键词： Graph embeddings

来源：评论

学校读者我要写书评

暂无评论

AKGIN: An API Knowledge Graph and Intent Network based Mashup-Oriented API Recommendation Method 26

AKGIN: An API Knowledge Graph and Intent Network based Mashu...

引用

26th International Conference on computer Supported Cooperative Work in Design, CSCWD 2023

作者： Yu, Changyong Hu, Rong Wang, Bangjun Hunan University of Science and Technology Computing and Novel Software Technology Hnust School of Computer Science and Engineering Xiangtan China

ISBN: (纸本)9798350331684

With the continuous development of the Web API ecosystem, mashup-oriented API recommendation gets a lot of attention. Collaborative filtering, deep learning and their combination based methods are recently proposed for API recommendation. However, the recommendation results often perform poorly when data is sparse. Therefore, we propose an API knowledge graph (AKG) and intent network based mashup-oriented API recommendation method, whose main target is to recruit the semantic representation ability to improve the precision of API recommendation. Firstly, Latent Dirichlet Allocation (LDA) is used to extract API topics, which are important entities in the AKG. Then, the mashup intention is modeled based on the mashup-API interactions. The high-order relationship paths are recursively aggregated into the representation vector of Mashup and API. Finally, the inner product of their representations present the probability of the API recommendation for the mashup. The experimental results on real-world datasets show that our proposed method outperforms several baseline methods in mashup-oriented API recommendations. © 2023 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Optimal Bounds for Distinct Quartics 51

Optimal Bounds for Distinct Quartics

引用

51st International Colloquium on Automata, Languages, and Programming, ICALP 2024

作者： Charalampopoulos, Panagiotis Gawrychowski, Pawel Ghazawi, Samah School of Computing and Mathematical Sciences Birkbeck University of London United Kingdom Institute of Computer Science University of Wroclaw Poland Department of Computer Science University of Haifa Israel Department of Software Engineering Braude College of Engineering Karmiel Israel

ISBN: (纸本)9783959773225

A fundamental concept related to strings is that of repetitions. It has been extensively studied in many versions, from both purely combinatorial and algorithmic angles. One of the most basic questions is how many distinct squares, i.e., distinct strings of the form UU, a string of length n can contain as fragments. It turns out that this is always O(n), and the bound cannot be improved to sublinear in n [Fraenkel and Simpson, JCTA 1998]. Several similar questions about repetitions in strings have been considered, and by now we seem to have a good understanding of their repetitive structure. For higher-dimensional strings, the basic concept of periodicity has been successfully extended and applied to design efficient algorithms – it is inherently more complex than for regular strings. Extending the notion of repetitions and understanding the repetitive structure of higher-dimensional strings is however far from complete. Quartics were introduced by Apostolico and Brimkov [TCS 2000] as analogues of squares in two dimensions. Charalampopoulos, Radoszewski, Rytter, Waleń, and Zuba [ESA 2020] proved that the number of distinct quartics in an n × n 2D string is O(n2 log2 n) and that they can be computed in O(n2 log2 n) time. Gawrychowski, Ghazawi, and Landau [SPIRE 2021] constructed an infinite family of n × n 2D strings with Ω(n2 log n) distinct quartics. This brings the challenge of determining asymptotically tight bounds. Here, we settle both the combinatorial and the algorithmic aspects of this question: the number of distinct quartics in an n × n 2D string is O(n2 log n) and they can be computed in the worst-case optimal O(n2 log n) time. As expected, our solution heavily exploits the periodic structure implied by occurrences of quartics. However, the two-dimensional nature of the problem introduces some technical challenges. Somewhat surprisingly, we overcome the final challenge for the combinatorial bound using a result of Marcus and Tardos [JCTA 2004] for permuta

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

FED-3DA: A Dynamic and Personalized Federated Learning Framework 48

FED-3DA: A Dynamic and Personalized Federated Learning Frame...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Wang, Hui Sun, Jie Wo, Tianyu Liu, Xudong Beihang University Sklsde School of Computer Science and Engineering Beijing China Beihang University Sklsde School of Software Beijing China

ISBN: (纸本)9781728163277

In federated learning, the non-IID data generated from heterogeneous clients may reduce the global model efficiency. Previous studies use personalization as a common approach to adapt the global model to these clients (called the local model). However, client's data distribution may change dynamically with its location or environment, which can degrade the performance of the local model, leading to a new Dynamic Personalized Federated Learning (DPFL) problem. This paper proposes a novel approach to reduce the impact of the dynamic distribution on the local model based on metalearning and distribution distance measurement named Fed-3DA. It calculates the distribution distance periodically to perceive the distribution change on the client and adjust the local model preferences from a global meta-model through the distribution representation. Our experiments on public datasets show that Fed-3DA can effectively reduce the performance fluctuation of the local model in DPFL scenarios. © 2023 IEEE.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

A novel transformer-based graph generation model for vectorized road design

A novel transformer-based graph generation model for vectori...

引用

作者： Zhou, Peichi Li, Chen Zhang, Jian Wang, Changbo Qin, Hong Liu, Long School of Software Engineering East China Normal University Shanghai China School of Computer Science and Technology East China Normal University Shanghai China Department of Computer Science StonyBrook University Stony BrookNY United States

Road network design, as an important part of landscape modeling, shows a great significance in automatic driving, video game development, and disaster simulation. To date, this task remains labor-intensive, tedious and time-consuming. Many improved techniques have been proposed during the last two decades. Nevertheless, most of the state-of-the-art methods still encounter problems of intuitiveness, usefulness and/or interactivity. As a rapid deviation from the conventional road design, this paper advocates an improved road modeling framework for automatic and interactive road production driven by geographical maps (including elevation, water, vegetation maps). Our method integrates the capability of flexible image generation models with powerful transformer architecture to afford a vectorized road network. We firstly construct a dataset that includes road graphs, density map and their corresponding geographical maps. Secondly, we develop a density map generation network based on image translation model with an attention mechanism to predict a road density map. The usage of density map facilitates faster convergence and better performance, which also serves as the input for road graph generation. Thirdly, we employ the transformer architecture to evolve density maps to road graphs. Our comprehensive experimental results have verified the efficiency, robustness and applicability of our newly-proposed framework for road design. © 2024 John Wiley & Sons Ltd.

关键词： Roads and streets

来源：评论

学校读者我要写书评

暂无评论

Learning dynamics of gradient descent optimization in deep neural networks

引用

science China(Information sciences) 2021年第5期64卷 17-31页

作者： Wei WU Xiaoyuan JING Wencai DU Guoliang CHEN School of Computer Science Wuhan University Institute of Data Science City University of Macau College of Computer Science and Software Engineering Shenzhen University

Stochastic gradient descent(SGD)-based optimizers play a key role in most deep learning models,yet the learning dynamics of the complex model remain obscure. SGD is the basic tool to optimize model parameters, and is improved in many derived forms including SGD momentum and Nesterov accelerated gradient(NAG). However, the learning dynamics of optimizer parameters have seldom been studied. We propose to understand the model dynamics from the perspective of control theory. We use the status transfer function to approximate parameter dynamics for different optimizers as the first-or second-order control system, thus explaining how the parameters theoretically affect the stability and convergence time of deep learning models, and verify our findings by numerical experiments.

关键词： learning dynamics deep neural networks gradient descent control model transferfunction

来源：评论

学校读者我要写书评

暂无评论

Can end-user feedback in social media be trusted for software evolution: Exploring and analyzing fake reviews

Can end-user feedback in social media be trusted for softwar...

引用

作者： Khan, Javed Ali Ullah, Tahir Khan, Arif Ali Yasin, Affan Akbar, Muhammad Azeem Aurangzeb, Khursheed Department of Computer Science Faculty of Physics Engineering and Computer Science University of Hertfordshire Hatfield United Kingdom School of Computer & Technology Beijing Institute of Technology Beijing China M3S Empirical Software Engineering Research Unit University of Oulu Oulu Finland School of Software Northwestern Polytechnical University Shaanxi Xian China Software Engineering Department Lappeenranta-Lahti University of Technology Lappeenranta Finland Department of Computer Engineering College of Computer and Information Sciences King Saud University Riyadh Saudi Arabia

End-user feedback in social media platforms, particularly in the app stores, is increasing exponentially with each passing day. software researchers and vendors started to mine end-user feedback by proposing text analytics methods and tools to extract useful information for software evolution and maintenance. In addition, research shows that positive feedback and high-star app ratings attract more users and increase downloads. However, it emerged in the fake review market, where software vendors started incorporating fake reviews against their corresponding applications to improve overall software ratings. For this purpose, we conducted an exploratory study to understand how end-users register and write fake reviews in the Google Play Store. We curated a research data set containing 68,000 end-user comments from the Google Play Store and a fake review generator, that is, the Testimonial generator (TG). Its purpose is to understand fake reviews on these platforms and identify the common patterns potential end-users and professionals use to report fake reviews by critically analyzing the end-user feedback. We conducted a detailed survey at the university of science and Technology Bannu, Pakistan, to identify the intelligence and accuracy of crowd-users in manually identifying fake reviews. In addition, we developed a ground truth to be compared with the results obtained from the automated machine and deep learning (M&DL) classifier experiment. In the survey, 512 end-users participated and recorded their responses in identifying fake reviews. Finally, various M&DL classifiers are employed to classify and identify end-user reviews into real and fake to automate the process. Unlike humans, the M&DL classifiers performed well in automatically classifying reviews into real and fake by obtaining much higher accuracy, precision, recall, and f-measures. The accuracy of manually identifying fake reviews by the crowd-users is 44.4%. In contrast, the M&DL classifiers obtained an

关键词： Application programs

来源：评论

学校读者我要写书评

暂无评论

MG-CR: Factor Memory Network and Graph Neural Network Based Personalized Course Recommendation 28th

MG-CR: Factor Memory Network and Graph Neural Network Based...

引用

28th International Conference on Database Systems for Advanced Applications, DASFAA 2023

作者： Zhang, Yun Yu, Minghe Sun, Jintong Zhang, Tiancheng Yu, Ge School of Computer Science and Engineering Northeastern University Shenyang China Software College Northeastern University Shenyang China

ISBN: (纸本)9783031306716

Course recommendations in universities select the most suitable courses for students according to their interests and academic requirements. However, existing works often focus on modeling course selection history, without paying sufficient attention to rich features contained in students’ personal information and courses’ teaching attributes, and various relationships between them. In order to realize the personalization and dynamics of course recommendation, we consider students and courses as two types of nodes to construct a heterogeneous information network (HIN), and propose a factor Memory network and Graph neural network based personalized Course Recommendation (MG-CR) on top of HIN. MG-CR captures multiple features of students and courses simultaneously in the input module, thereby maximizing the retention of personalized information. The feature transfer module optimizes course features by considering various types of relationships between courses in the HIN, and together with the input module, it ensures high recommendation accuracy in the cold-start scenario. The updating module explicitly stores the student’s interest level for each factor in the state matrix, which is then used to dynamically recommend courses for students in the predicting module. Experimental results show that MG-CR provides accurate, dynamic and personalized recommendations, and has superior performance in the cold-start scenario. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Sentiment analysis in social internet of things using contextual representations and dilated convolution neural network

引用

Neural Computing and Applications 2024年第20期36卷 12357-12370页

作者： Abid, Fazeel Rasheed, Jawad Hamdi, Mohammed Alshahrani, Hani Al Reshan, Mana Saleh Shaikh, Asadullah Department of Information Systems Dr. Hassan Murad School of Management University of Management and Technology Lahore Pakistan Department of Computer Engineering Istanbul Sabahattin Zaim University Istanbul34303 Turkey Department of Software Engineering Istanbul Nisantasi University Istanbul34394 Turkey Department Computer Science College of Computer Science and Information Systems Najran University Najran61441 Saudi Arabia Department Information Systems College of Computer Science and Information Systems Najran University Najran61441 Saudi Arabia

The methodologies based on neural networks are substantial to accomplish sentiment analysis in the Social Internet of Things (SIoT). With social media sentiment analysis, significant insights can produce efficient and intelligent applications. Neural networks such as recurrent neural networks (RNNs) and convolution neural networks (CNNs) have been considered widely in many text classification tasks. However, RNNs are computationally expensive and require complex training to capture contextual information and long-term dependencies. Similarly, traditional CNNs must stack multiple convolutional layers, requiring massive computations and additional parameters. To address these problems, this work initialized the novel architecture, in which contextual representations (CRs) based on the textual framework are proposed at the initial step. In CRs, state-of-the-art word representation models, such as GloVe (global vectors) and FastText (subword information), collectively produce word representations upon the input sequence using a weight mechanism. Secondly, a unique way is introduced: a three-parallel layered dilated convolutional network with global mean pooling. The experimental results show that the proposed methods when compared with baseline methods, the dilation in CNNs following CRs significantly increases the accuracy from 72.45 to 98.98% and reduces computational resources. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Adaptive IoT Decision Making in Uncertain Environments 7

Adaptive IoT Decision Making in Uncertain Environments

引用

7th IEEE International Conference on Smart Internet of Things, SmartIoT 2023

作者： Wang, Yan Zheng, Liwei He, Jingjing Cui, Zhanqi Computer School Software Engineering Research Center Beijing Information Science and Technology University Beijing China

ISBN: (纸本)9798350316575

The rapid growth of the Internet of Things (IoT) has led to widespread deployment of IoT systems in domains such as smart homes, healthcare, and transportation. However, IoT systems often operate under uncertainty, making it difficult to predict and control their behavior. In this paper, we propose an adaptive decision making approach for IoT systems in uncertain, dynamic environments. We present a framework with perception, decision and execution layers to handle uncertainty in IoT systems. The perception layer senses the environment and system state. The decision layer employs an optimized deep Q-network algorithm (Ad-DQN) specifically designed to handle uncertain environments, enabling it to make informed decisions based on learned experiences. The execution layer implements the actions. We demonstrate the framework on an intelligent air conditioning system as a case study of an IoT system operating under uncertainty. The Ad-DQN based decision layer adapts the air conditioning control policy to maximize comfort while minimizing energy usage. Experiments show our method outperforms traditional DQN method in uncertain environments. © 2023 IEEE.

关键词： Reinforcement learning

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：