检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Stigall, William A. College of Computing and Software Engineering Kennesaw State University Marietta United States

In this study we investigate the performance of Deep Q-Networks utilizing Convolutional Neural Networks (CNNs) and Transformer architectures across 3 different Atari Games. The advent of DQNs have significantly advanced Reinforcement Learning, enabling agents to directly learn optimal policy from high dimensional sensory inputs from pixel or RAM data. While CNN based DQNs have been extensively studied and deployed in various domains Transformer based DQNs are relatively unexplored. Our research aims to fill this gap by benchmarking the performance of both DCQNs and DTQNs across the Atari games’ Asteroids, SpaceInvaders and Centipede. We find that in the 35-40 million parameter range, the DCQN outperforms the DTQN in speed across both ViT and Projection Architectures, We also find the DCQN outperforms the DTQN in all games except for centipede. © 2024, CC BY.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Phishing Website Detection with and Without Proper Feature Selection Techniques: Machine Learning Approach

Phishing Website Detection with and Without Proper Feature S...

引用

作者： Adane, Kibreab Beyene, Berhanu Faculty of Computing and Software Engineering Arba Minch University Arba Minch Ethiopia Computer Science Department Ethiopian Civil Service University Addis Ababa Ethiopia

Indeed, successful phishing website attempts could result in catastrophic data loss, login credential compromise, ransomware infection, and financial loss. It also significantly hampers the competitiveness and productivity of online users and internet-dependent organizations unless an intelligent anti-phishing solution is devised. Due to detecting fresh phishing website attacks with maximum accuracy by discovering hidden patterns from complex datasets is shown to be an intrinsic property of M- Learning approaches, the study conducted rigorous experiments on four purposely selected efficient supervised M-Learning algorithms before and after applying five widely used proper feature selection techniques such as Recursive Feature Elimination, Pierson Correlation Coefficient, Principal Component Analysis, Uni-variate Feature Selection, and Mutual Information. The proposed study was conducted to balance the research gaps and scientific disputes in the rigorously reviewed studies. The study’s final outcome is a proposal for an intelligent phishing website model that yields higher accuracy, faster response times, and fewer average misclassification rates. The study also explored the feature selection techniques that had more, less, and no contributions to enhancing each classifier's accuracy. As compared to the remaining classifiers, the Cat-Boost Classifier attained superior phishing website detection accuracy (97.46%), F1-score (97.49%), a lower average misclassification rate (2.54%), and acceptable train-test computational time (7 s) after using the UFS technique. On the other hand, the PCA technique failed to enhance the accuracy of the Cat-Boost, Gradient-Boost, and Random Forest Classifiers due to scoring less accuracy than the accuracy reached before using proper feature selection techniques. To obtain more promising results, in future work, phishing website detection is expected to be carried out using a Hybrid proper feature selection technique, huge datasets, prop

关键词： Feature Selection

来源：评论

学校读者我要写书评

暂无评论

Adaptive Occlusion Face Recognition Feature Fusion Network Based on Precise Guidance Mask Extraction

引用

Journal of Shanghai Jiaotong University (Science) 2025年 1-12页

作者： Ye, Jihua Zhang, Lei Wang, Chao Tu, Qixuan Zheng, Rehua Huang, Huiyuan Jiang, Aiwen School of Computer and Information Engineering Jiangxi Normal University Nanchang330022 China Jiangxi Provincial Key Laboratory for High Performance Computing Jiangxi Normal University Nanchang330022 China State International Science & Technology Cooperation Base of Networked Supporting Software Jiangxi Normal University Nanchang330022 China

The uncertainty in the position and size of occluding objects greatly affects the extraction of identity features in facial recognition, which is a challenge that existing methods fail to effectively address. To tackle this issue, this paper introduces a novel feature fusion network for occluded facial recognition that adaptively handles arbitrary occlusions and accurately extracts occlusion information masks. The network effectively mitigates the impact of occlusions on facial identity features through a feature fusion module. The proposed method employs an occlusion classification module and an occlusion classification loss to precisely guide the extraction of accurate occlusion masks. Subsequently, an occlusion-aware hybrid attention module is introduced, which focuses on the relationship between facial features and occlusion masks across channel and spatial dimensions. This fusion strategy enables the extraction of facial identity features that are resilient to occlusions. Experimental results indicate that the proposed method holds a significant advantage in handling facial recognition tasks with arbitrary occlusions. On the LFW and MF1 synthetic occlusion datasets, the average accuracy rates reached 93.19% and 72.57%, respectively, showing improvements of 1.7 and 2.39 percent points over mainstream methods. © Shanghai Jiao Tong University 2025.

关键词： A feature fusion occluded facial recognition occlusion mask occlusion perception TP391.41

来源：评论

学校读者我要写书评

暂无评论

Age-Based Differences in Drone Control Gestures: An Exploratory Study 35

Age-Based Differences in Drone Control Gestures: An Explorat...

引用

35th Australian Computer-Human Interaction Conference, OzCHI 2023

作者： Hermann, Julia Dogangün, Aysegül Hesenius, Marc Institute of Positive Computing Ruhr West University of Applied Sciences Bottrop46236 Germany Institute for Software Engineering University of Duisburg-Essen Essen45127 Germany

ISBN: (纸本)9798400717079

In addition to cultural background and professional context, age can also influence user preferences for gestures when controlling drones. In this paper, we present an exploratory study of age-based differences in gesture selection and design for drone control with 24 participants aged 18 to 69. We investigated the impact of age on the shape, type, body parts used, and number of partial gestures employed for drone direction and interaction commands. Firstly, we quantified the number and type of gestures utilized by each age group for specific commands, revealing significant disparities between older and younger participants. Additionally, we explore the qualitative aspects of gesture control by analyzing participants' descriptions and verbalized thoughts, uncovering further differences and similarities. Overall, our paper contributes to the growing body of research on Human-Drone Interaction by highlighting the significance of age in gesture control and its consequences for the individual User Experience. By incorporating these findings into future design practices, we can advance the field and promote inclusive drone interfaces that accommodate users of all ages. © 2023 Owner/Author.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Stock Price Prediction using CuDNNLSTM and multiple CNN layers

Stock Price Prediction using CuDNNLSTM and multiple CNN laye...

引用

2023 IEEE International Conference on Artificial Intelligence, Blockchain, and Internet of Things, AIBThings 2023

作者： Kanwal, Anika Chandrasekaran, Siva Akhunzada, Adnan Swinburne University of Technology Department of Computer Science and Software Engineering Melbourne Australia University of Doha for Science and Technology Department of Computing and Information Technology Doha Qatar

ISBN: (纸本)9798350322347

Academic and financial sectors are interested in research areas that focus on understanding the patterns of financial activities and predicting their future changes. The daily movement of financial data involves complex, insufficient, and ambiguous information, making its prediction an extremely difficult endeavour. The issues of forecasting and analysing financial data is therefore complex and time-dependent. Deep neural networks (DNNs) can be used to solve nonlinear problems more effectively than conventional machine learning techniques. This study proposes an intelligent and optimal model CuDNNLSTM-Multi(1dCNN) for stock market price prediction using a hybrid of CUDA based Long Short Term Memory (LSTM) and Multiple One Dimensional Convolution Neural Networks (CNNs). A variety of models, including Multiple CuDNNLSTM-CNN, Multi-CuDNNLSTM, Multi (1dCNN), are used to compare the prediction performance of the proposed model. In terms of RMSE, MAE, and MAPE, the results show that the suggested model outperforms all hybrid and individual models. © 2023 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

A Rational Decision-Making Approach Based on Bayesian Network and BDI Model For Autonomous Driving System

A Rational Decision-Making Approach Based on Bayesian Networ...

引用

2023 International Conference on Computer Network Security and software engineering, CNSSE 2023

作者： Zhang, Xinyuan Du, Dehui Software Engineering Institute East China Ormal University Shanghai200062 China Shanghai Key Laboratory of Trnstworthy Computing Shanghai200062 China Intemational Joint Laboratory of Trnstworthy Software of Ministry of Education Shanghai200062 China

ISBN: (数字)9781510666528

ISBN: (纸本)9781510666511

The safety of decision-making in autonomous driving systems (ADSs) is a challenging issue, which is very important for ADS development. As a highly acceptable decision method, Bayesian Net\vork (BN) has attracted more and more attention, therefore, its decision confidence and robustness have also become a focus. Since BN is a form of reasoning based on probability, for the decision results with low confidence (that is, the probability values of several optional decision actions have a small difference), when the data is slightly disturbed, it is likely to change, resulting in serious consequences such as car crash. To generate safe decision-making, we innovatively propose a rational Bayesian network decision-making approach based on BDI model. It helps to improve the decision confidence of the traditional B. BDI model is a well-known theory of inferencing agents' mental states (belief, desire and intention). We use it to guide the decision-making process of Bayesian network. According to the domain knowledge of ADS, we introduce and design rule-based intention inference for the decision agent to build a Bayesian net\vork with BDI-layer. And for other uncertain agents in the environment, we utilize LSTM model to predict their intentions and provide scenario information for the construction of the above network. To sum up, we propose a rational decision-making approach based on Bayesian network guided by BDI model. Our novel approach makes the traditional Bayesian network decision more humanized and improves the confidence of decision results. Finally, we take the lane-change and overtaking scenario as an example to illustrate our approach in detail and demonstrate the effectiveness in improving decision confidence. © 2023 SPIE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Molecular Dynamics and Protein-Protein Interactions (PPIs) on SARS-CoV-2 coronavirus variants 24

Molecular Dynamics and Protein-Protein Interactions (PPIs) o...

引用

24th Annual Conference on Information Technology Education, SIGITE 2023

作者： Addepalli, Durga Narayana Varma Chen, Jiawei Xie, Chloe Yixin College of Computing and Software Engineering Kennesaw State University MariettaGA United States College of Letter and Science University of California BerkeleyCA United States

来源：评论

学校读者我要写书评

暂无评论

Research on global register allocation for code containing array-unit dual-usage register names

Research on global register allocation for code containing a...

引用

作者： Hu, Yonghua Zhang, Xin Wang, Shuying Liang, Wei Li, Kuan-Ching School of Computer Science and Engineering Hunan University of Science and Technology Xiangtan China Hunan Key Laboratory for Service Computing and Novel Software Technology Xiangtan China

Array-unit dual-usage register is a kind of register resource that can be read or written as a whole or individually. It is mainly configured in processors with SIMD processing units and provides register-level speed data transfer between the scalar and vector processing units. To improve the efficiency of algorithms by using an array-unit dual-usage register, we investigate in this article the problem of adapting register allocation to code containing array-unit dual-usage register names. We propose a corresponding global register allocation method by combining the allocation of regular registers with array-unit dual-usage register, ensuring that the names of array-unit dual-usage register can be used in the input code of register allocation. Moreover, we present the processing framework of this method and the specific algorithms of some related vital aspects and demonstrate the working principles of the algorithms by an example. Experimental studies were conducted on a platform based on the FT-M7002 DSP core, and showing that our register allocation method can effectively handle codes containing array-unit dual-usage register names and support relevant application algorithms to improve their data transfer scheme. For some typical algorithms with input matrix, substantial performance improvements of twofold or higher are achieved. © 2022 John Wiley & Sons, Ltd.

关键词： Data transfer

来源：评论

学校读者我要写书评

暂无评论

Accurate Reliability Analysis Methods for Approximate computing Circuits

引用

Tsinghua Science and Technology 2022年第4期27卷 729-740页

作者： Zhen Wang Guofa Zhang Jing Ye Jianhui Jiang Fengyong Li Yong Wang School of the Computer Science and Technology Shanghai University of Electric PowerShanghai 200090China Institute of Computing Technology Chinese Academy of SciencesBeijing 100190China School of Software Engineering Tongji UniversityShanghai 200092China

In recent years,Approximate computing Circuits(ACCs)have been widely used in applications with intrinsic tolerance to *** the increased availability of approximate computing circuit approaches,reliability analysis methods for assessing their fault vulnerability have become highly *** this study,two accurate reliability evaluation methods for approximate computing circuits are *** reliability of approximate computing circuits is calculated on the basis of the iterative Probabilistic Transfer Matrix(PTM)*** the calculation,the correlation coefficients are derived and combined to deal with the correlation problem caused by fanout *** accuracy and scalability of the two methods are verified using three sets of approximate computing circuit instances and more circuits in Evo Approx8 b,which is an approximate computing circuit open source *** results show that relative to the Monte Carlo simulation,the two methods achieve average error rates of 0.46%and 1.29%and time overheads of 0.002%and 0.1%.Different from the existing approaches to reliability estimation for approximate computing circuits based on the original PTM model,the proposed methods reduce the space overheads by nearly 50%and achieve time overheads of 1.78%and 2.19%.

关键词： Approximate computing Circuit(ACC) correlation coefficient iterative Probabilistic Transfer Matrix(PTM) reliability

来源：评论

学校读者我要写书评

暂无评论

Leaching Fraction (LF) of Irrigation Water for Saline Soils Using Machine Learning

引用

Intelligent Automation & Soft computing 2023年第5期36卷 1915-1930页

作者： Rab Nawaz Bashir Imran Sarwar Bajwa Muhammad Waseem Iqbal Muhammad Usman Ashraf Ahmed Mohammed Alghamdi Adel ABahaddad Khalid Ali Almarhabi Department of Computer Science COMSATS University IslamabadVehari CampusPakistan Department of Computer Science and IT Islamia University BahawalpurPakistan Department of Software Engineering Superior University Lahore54000Pakistan Department of Computer Science GC Women UniversitySialkot53310Pakistan Department of Software Engineering College of Computer Science and EngineeringUniversity of Jeddah21493Saudi Arabia Department of Information System Faculty of Computing and Information TechnologyKing Abdulaziz UniversityJeddah21589Saudi Arabia Department of Computer Science College of Computing in Al-QunfudahUmm Al-Qura UniversityMakkah24381Saudi Arabia

Soil salinity is a serious land degradation issue in *** is a major threat to agriculture *** irrigation water is applied to leach down the salts from the root zone of the plants in the form of a Leaching fraction(LF)of irrigation *** the leaching process to be effective,the LF of irriga-tion water needs to be adjusted according to the environmental conditions and soil salinity level in the form of Evapotranspiration(ET)*** relationship between environmental conditions and ET rate is hard to be deﬁned by a linear relationship and data-driven Machine learning(ML)based decisions are required to determine the calibrated Evapotranspiration(ETc)***-assisted ETc is pro-posed to adjust the LF according to the ETc and soil salinity level.A regression model is proposed to determine the ETc rate according to the prevailing tempera-ture,humidity,and sunshine,which would be used to determine the smart LF according to the ETc and soil salinity *** proposed model is trained and tested against the Blaney Criddle method of Reference evapotranspiration(ETo)*** validation of the model from the test dataset reveals the accu-racy of the ML model in terms of Root mean squared errors(RMSE)are 0.41,Mean absolute errors(MAE)are 0.34,and Mean squared errors(MSE)are 0.28 mm *** applications of the proposed solution in a real-time environ-ment show that the LF by the proposed solution is more effective in reducing the soil salinity as compared to the traditional process of leaching.

关键词： Leaching fraction saline soil evapotranspiration machine learning calibrated evapotranspiration artiﬁcial intelligence blaney criddle method

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：