检索结果-内蒙古大学图书馆

International Joint Conference on Neural Networks (IJCNN)

作者： Shaojie Zhang Yiwei Ding Enrui Hu Yue Yu Yu Zhang Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China Peng Cheng Laboratory Shenzhen China Distributed and Parallel Software Lab Huawei Shenzhen China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Code representation learning is an important way to encode the semantics of source code through pre-training. The learned representation supports a variety of downstream tasks, such as natural language code search and code defect detection. Inspired by pre-trained models for natural language representation learning, existing approaches often treat the source code or its structural information (e.g., Abstract Syntax Tree or AST) as a plain token sequence. Unlike natural language, programming language has its unique code unit information (e.g., identifiers and expressions) and logic information (e.g., the functionality of a code snippet). To further explore those properties, we propose Abstract Code Embedding (AbCE), a self-supervised learning method that considers the abstract semantics of code logic. Instead of scattered tokens, AbCE treats an entire node or a subtree in an AST as a basic code unit during pre-training, which preserves the entirety of a coding unit. Moreover, AbCE learns the abstract semantics of AST nodes via a self-distillation way. Experimental results show that it achieves significant improvements over state-of-the-art baselines on code search tasks and comparable performance on code clone detection and defect detection tasks even without using contrastive learning or curriculum learning.

关键词： Representation learning Codes Source coding Semantics Natural languages Cloning Syntactics

来源：评论

学校读者我要写书评

暂无评论

Hybrid Evolutionary Algorithm Based Relevance Feedback Approach for Image Retrieval

引用

computers, Materials & Continua 2022年第1期70卷 963-979页

作者： Awais Mahmood Muhammad Imran Aun Irtaza Qammar Abbas Habib Dhahri Esam Mohammed Asem Othman Arif Jamal Malik Aaqif Afzaal Abbasi College of Applied Computer Science King Saud University(Almuzahmiyah Campus)Riyadh11543Saudi Arabia Shaheed Zulfikar Ali Bhutto Institute of Science and Technology(Islamabad Campus) 44000Pakistan University of Engineering and Technology Taxila Pakistan International Islamic University Islamabad44000Pakistan Faculty of Sciences and Technology University of KairouanSidi Bouzid4352Tunisia Department of Software Engineering Foundation University IslamabadIslamabad44000Pakistan

Searching images fromthe large image databases is one of the potential research areas of multimedia *** most challenging task for nay CBIR system is to capture the high level semantic of *** researchers of multimedia domain are trying to fix this issue with the help of Relevance Feedback(RF).However existing RF based approaches needs a number of iteration to fulfill user’s *** paper proposed a novel methodology to achieve better results in early iteration to reduce the user interaction with the *** previous research work it is reported that SVM based RF approach generating better results for ***,this paper focused on SVM based RF *** enhance the performance of SVM based RF approach this research work applied Particle Swarm Optimization(PSO)and Genetic Algorithm(GA)before applying SVM on user *** main objective of using thesemeta-heuristic was to increase the positive image sample size from *** steps PSO is applied by incorporating the user feedback and secondly GA is applied on the result generated through PSO,finally SVM is applied using the positive sample generated through *** proposed technique is named as Particle Swarm Optimization Genetic Algorithm-Support Vector Machine Relevance Feedback(PSO-G A-SVMRF).Precisions,recall and F-score are used as performance metrics for the assessment and validation of PSO-GA-SVM-RF approach and experiments are conducted on coral image dataset having 10908 *** experimental results it is proved that PSO-GA-SVM-RF approach outperformed then various well known CBIR approaches.

关键词： Feature selection image retrieval particle swarm optimization

来源：评论

学校读者我要写书评

暂无评论

Development of Area-Based Explosive Trace Detection Using Deep Transfer Learning

Development of Area-Based Explosive Trace Detection Using De...

引用

Innovative engineering sciences and Technological Research (ICIESTR), International Conference on

作者： Monday F. Ohemu Ambrose A. Azeta Ibrahim A. Adeyanju Chukwuemeka C. Obasi Department of Electrical/Electronics Engineering Airforce Institute of Technology Kaduna Nigeria Department of Software Engineering Namibia University of Science and Technology Namibia Department of Computer Engineering Federal University Oye-Ekiti Nigeria Department of Computer Engineering Edo State University Uzairue

ISBN: (数字)9798350348637

ISBN: (纸本)9798350348644

Terrorism poses a significant global threat, necessitating comprehensive security measures to protect people and critical infrastructure. With the increasing frequency of explosive-based attacks, there is a pressing need for advanced detection systems to mitigate potential devastation. This paper presents the development of an area-based explosive trace detection system using deep transfer learning. Leveraging deep learning technology trained on a substantial dataset collected from sensor networks, the proposed model, named Deep Transfer Learning for Explosive Trace Detection (DTLETD), demonstrates remarkable capabilities in classifying explosive gases based on concentrations of Carbon (C), Hydrogen (H), Oxygen (O), and Nitrogen (N). The methodology involves converting the dataset into 2D data using a serial data to image generator, followed by the adaptation of a pre-trained model for transfer learning. The DTLETD model outperformed conventional Convolutional Neural Network (CNN) models, achieving a training time reduction of about 92 seconds compared to 1287 seconds for the CNN model. Additionally, the transfer learning model exhibits faster convergence with nearly zero losses during training and validation, yielding an impressive accuracy of 99.7% and an average AUC value of approximately 0.89. This research significantly advances the field of explosive trace detection by integrating machine learning-based approaches with deep transfer learning techniques, thereby enhancing security protocols and mitigating potential threats.

关键词： Training Adaptation models Accuracy Protocols Terrorism Transfer learning Pressing Explosives Data models Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Learning and Motivational Impact of Game-Based Learning: Comparing Face-To-Face and Online Formats on computer science Education

arXiv

引用

arXiv 2024年

作者： López-Fernández, Daniel Gordillo, Aldo Pérez, Jennifer Tovar, Edmundo Computer Science Department Universidad Politécnica de Madrid Madrid Spain Computer Languages and Systems and Software Engineering Department Universidad Politécnica de Madrid Madrid Spain

Contribution: This article analyzes learning and motivational impact of teacher-authored educational video games on computer science education and compares its effectiveness in both face-to-face and online (remote) formats. This work presents comparative data and findings obtained from 217 students who played the game in a face-to-face format (control group) and 104 students who played the game in an online format (experimental group). Background: Serious video games have been proven effective at computer science education, however it is still unknown whether the effectiveness of these games is the same regardless of their format, face-to-face or online. Moreover, the usage of games created through authoring tools has barely been explored. Research questions: Are teacher-authored educational video games effective in terms of learning and motivation for computer science students? Does the effectiveness of teacher-authored educational video games depend on whether they are used in a face-to-face or online format? Methodology: A quasi-experiment has been conducted by using three instruments (pre-test, post-test and questionnaire) with the purpose of comparing the effectiveness of Game-Based Learning in face-to-face and online formats. A total of 321 computer science students played a teacher-authored educational video game aimed to learn about software design. Findings: The results reveal that teacher-authored educational video games are highly effective in terms of knowledge acquisition and motivation both in face-to-face and online formats. The results also show that some students’ perceptions were more positive when a face-to-face format was used. Copyright © 2024, The Authors. All rights reserved.

关键词： Motivation

来源：评论

学校读者我要写书评

暂无评论

An Effective Post - Filtering for Minimum Variance Distortionless Response Beamformer

An Effective Post - Filtering for Minimum Variance Distortio...

引用

2023 Seminar on Signal Processing, SoSP 2023

作者： The, Quan Trong Huy, Nguyen Ba Anh, Pham Tuan Ha Noi University of Science and Technology Falculty of Information Technology Ha Noi Viet Nam University ITMO Falculty of Control Systems and Robotics Saint-Petersburg Russia University ITMO Falculty of Software Engineering and Computer Systems Saint-Petersburg Russia

ISBN: (纸本)9798350371086

An effective hands-free speech interface with an appropriate digital signal processing is essential for teleconference systems and control systems. Beamforming technique, which use a microphone array (MA) is an ideal solution for distant - talking speech acquisition, is one of the most attractive approaches for suppressing background noise and improving the speech intelligibility and speech quality of the processing system. Minimum Variance Distortionless Response (MVDR) has the capacity of maximizing the sensitivity in a certain direction at aiming to extract the desired target speech while removing all total power of surrounding noise. However, in many complex recording environments, due to the error of direction of arrival of interest speaker, the difference of microphone sensitivities, imprecise designed geometry of MA, the performance of MVDR beamformer often degraded. In this contribution, the authors suggested an additive post-filtering to enhance the effectiveness of MVDR beamformer. © 2023 IEEE.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

A Parallel Neighborhood Search Algorithm for Hybrid Disassembly Line Balancing Problem Considering Disabled Workers 19

A Parallel Neighborhood Search Algorithm for Hybrid Disassem...

引用

19th IEEE International Conference on Networking, Sensing and Control, ICNSC 2022

作者： Dou, Zheng Liu, Peisheng Guo, XiWang Wang, Jiacun Qin, Shujin Qi, Liang Liaoning Petrochemical University Information and Control Fushun China Liaoning Petrochemical University Artificial Intelligence and Software Fushun China Liaoning Petrochemical University Monmouth University Information and Control United States Monmouth University Computer Science and Software Engineering United States Shangqiu Normal University Economics and Management Shangqiu China Shandong University of Science and Technology Computer Intelligent Science and Technology Qingdao China

ISBN: (纸本)9781665472432

The recycling and reuse of waste products is increasingly inseparable from disassembly lines. This study addresses a multi-product hybrid disassembly line balancing problem that is composed of a U-shaped line and a single-row layout. Disabled workers are considered. According to the characteristics of the problem, a mathematical model for maximizing the recovery profit is established. Combined with the actual disassembly process and considering the limitations of the disabled workers in product selection, two types of neighborhood structures are designed using a parallel neighborhood search algorithm (PNSA) to find feasible solutions. Experimental analysis shows that the model and algorithm proposed in this paper effectively solve the above problems. © 2022 IEEE.

关键词： Optimization

来源：评论

学校读者我要写书评

暂无评论

Adaptive Graph Structure Learning Neural Rough Differential Equations for Multivariate Time Series Forecasting

引用

IEEE Transactions on Big Data 2025年

作者： Su, Yuming Ma, Tinghuai Rong, Huan Wahab, Mohamed Magdy Abdel Nanjing University of Information Science & Technology The School of Cyber Science and Engineering School of Computer Science Nanjing210044 China Jiangsu Ocean University School of Computer Engineering Lianyungang222005 China Nanjing University of Information Science & Technology School of Software Nanjing210044 China Nanjing University of Information Science & Technology School of Artificial Intelligence School of Future Technology Nanjing210044 China Cairo University Faculty of Science Cario 12613 Egypt

Multivariate time series forecasting has extensive applications in urban computing, such as financial analysis, weather prediction, and traffic forecasting. Using graph structures to model the complex correlations among variables in time series, and leveraging graph neural networks and recurrent neural networks for temporal aggregation and spatial propagation stage, has shown promise. However, traditional methods' graph structure node learning and discrete neural architecture are not sensitive to issues such as sudden changes, time variance, and irregular sampling often found in real-world data. To address these challenges, we propose a method called Adaptive Graph structure Learning neural Rough Differential Equations (AGLRDE). Specifically, we combine dynamic and static graph structure learning to adaptively generate a more robust graph representation. Then we employ a spatio-temporal encoderdecoder based on Neural Rough Differential Equations (Neural RDE) to model spatio-temporal dependencies. Additionally, we introduce a path reconstruction loss to constrain the path generation stage. We conduct experiments on six benchmark datasets, demonstrating that our proposed method outperforms existing state-of-the-art methods. The results show that AGLRDE effectively handles aforementioned challenges, significantly improving the accuracy of multivariate time series forecasting. © 2015 IEEE.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

A Context-Driven Approach for Co-Auditing Smart Contracts Using Gpt-4 Code Interpreter

SSRN

引用

SSRN 2024年

作者： Bouafif, Mohamed Salah Zheng, Chen Qasse, Ilham Zulkoski, Ed Hamdaqa, Mohammad Khomh, Foutse Department of Computer and Software Engineering Polytechnique Montreal Montreal Canada Department of Computer Science Reykjavik University Reykjavík Iceland Department of Computer Science The Hong Kong Polytechnic University Hung Hom Hong Kong Quantstamp Dover United States

Context: The surge in the adoption of smart contracts necessitates rigorous auditing to ensure their security and reliability. Manual auditing, although comprehensive, is time-consuming and heavily reliant on the auditor's expertise. With the rise of Large Language Models (LLMs), there is growing interest in leveraging them to assist auditors in the auditing process (co-auditing). However, the effectiveness of LLMs in smart contract co-auditing is contingent upon the design of the input prompts, especially in terms of context description and code ***: This paper introduces a novel context-driven prompting technique for smart contract co-auditing. The primary objective is to improve the effectiveness of LLMs in detecting vulnerabilities in smart contracts through enhanced prompt ***: Our approach employs three techniques for context scoping and augmentation, namely: Code Scoping, which involves chunking long code into self-contained code segments based on code interdependencies;Assessment Scoping, which enhances context description based on the target assessment goal to limit the search space;and Reporting Scoping, which forces a specific format for the generated ***: Through empirical evaluations on publicly available vulnerable contracts, our method demonstrated a detection rate of 96\% for vulnerable functions, outperforming the native prompting approach, which detected only 53\%. To assess the reliability of our prompting approach, manual analysis of the results was conducted by two expert auditors from Quantstamp, a world-leading smart contract auditing company. The experts' analysis indicates that, in unlabeled datasets, our proposed approach enhances the proficiency of the GPT-4 code interpreter in detecting ***: The context-driven prompting technique significantly improves the effectiveness of LLMs in smart contract co-auditing. This method not only enhances the detection rate of vulnerabilities but

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Recognition of Handwritten Tamazight Characters Using ResNet, MobileNet and VGG Transfer Learning 23

Recognition of Handwritten Tamazight Characters Using ResNet...

引用

23rd International Conference on New Trends in Intelligent software Methodologies, Tools and Techniques, SoMeT 2024

作者： Fattahi, Jaouhar Sghaier, Feriel Mejri, Mohamed Manai, Elyes Ghayoula, Ridha Department of Computer Science and Software Engineering Laval University 2325 rue de l'Université QuébecQCG1V 0A6 Canada Ecole Nationale d'Ingénieurs de Carthage Université de Carthage Tunis Tunisia Faculty of Engineering Information Systems Department University of Moncton New Brunswick Canada

ISBN: (纸本)9781643685380

The Tamazight civilization stands as a significant cultural entity, marked by its linguistic diversity, historical legacy, and scriptural traditions, which collectively enrich the cultural tapestry of North Africa. Among these traditions, the Tamazight handwritten script assumes particular importance, embodying centuries of cultural identity and artistic expression. Recognizing the imperative of safeguarding this cultural heritage, our study focuses on Tamazight handwritten character recognition. Leveraging the strategic application of Transfer Learning, we explore its efficacy in this domain. Transfer Learning presents a robust framework wherein pre-existing models are adapted for specific tasks despite limited data availability. Our research employs three prominent Transfer Learning architectures: VGG, ResNet, and MobileNet. Through a rigorous comparative analysis, we discern the efficacy of these methodologies in the context of Tamazight handwritten character recognition. Our findings underscore the potential of Transfer Learning to significantly augment the accuracy and efficiency of script recognition systems, thereby advancing the overarching objective of preserving and propagating the Tamazight cultural heritage. © 2024 IOS Press. All rights reserved.

关键词： Transfer learning

来源：评论

学校读者我要写书评

暂无评论

GRL-Prompt: Towards Knowledge Graph based Prompt Optimization via Reinforcement Learning

arXiv

引用

arXiv 2024年

作者： Liu, Yuze Liu, Tingjie Zhang, Tiehua Xia, Youhua Wang, Jinze Shen, Zhishu Jin, Jiong Yu, Fei Richard Ant Group Shanghai China Shenzhen China School of Computer Science and Technology Tongji University Shanghai China Computing and Engineering Technologies Swinburne University of Technology Melbourne Australia School of Computer Science and Artificial Intelligence Wuhan University of Technology Wuhan China College of Computer Science and Software Engineering Shenzhen University Shenzhen China

Large language models (LLMs) have demonstrated impressive success in a wide range of natural language processing (NLP) tasks due to their extensive general knowledge of the world. Recent works discovered that the performance of LLMs is heavily dependent on the input prompt. However, prompt engineering is usually done manually in a trial-and-error fashion, which can be labor-intensive and challenging in order to find the optimal prompts. To address these problems and unleash the utmost potential of LLMs, we propose a novel LLMsagnostic framework for prompt optimization, namely GRL-Prompt, which aims to automatically construct optimal prompts via reinforcement learning (RL) in an end-to-end manner. To provide structured action/state representation for optimizing prompts, we construct a knowledge graph (KG) that better encodes the correlation between the user query and candidate in-context examples. Furthermore, a policy network is formulated to generate the optimal action by selecting a set of in-context examples in a rewardable order to construct the prompt. Additionally, the embedding-based reward shaping is utilized to stabilize the RL training process. The experimental results show that GRL-Prompt outperforms recent state-ofthe- art methods, achieving an average increase of 0.10 in ROUGE-1, 0.07 in ROUGE-2, 0.07 in ROUGE-L, and 0.05 in BLEU. Copyright © 2024, The Authors. All rights reserved.

关键词： Knowledge graph

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：