检索结果-内蒙古大学图书馆

Holography, Diffractive Optics, and Applications XIV 2024

作者： Wang, Pin Fu, Qingyang Zhang, Yaping Poon, Ting-Chung Yunnan Provincial Key Laboratory of Modern Information Optics Kunming University of Science and Technology Kunming Yunnan650500 China Nanjing University of Science and Technology Nanjing Jiangsu210094 China Bradley Department of Electrical and Computer Engineering Virginia Tech BlacksburgVA24061 United States

ISBN: (纸本)9781510682085

The fully analytical method has been widely studied for its higher accuracy and efficiency, and the addition of shading rendering can effectively improve the three-dimensional perception of reconstructed images. However, existing fully analytical polygon-based shading rendering methods suffer from high computational complexity and complexity due to the need for additional calculation of the highlight spectrum. Based on previous research, we employ a continuous light and shadow algorithm that uses linear interpolation to decouple the reflection term from the spectrum, meaning that we incorporate the specular component into the calculation of vertex modulus, and separately compute the vertex modulus and the corresponding spectrum values, simplifying the calculation process of existing Blinn Phong lighting models. The algorithm can effectively generate highly realistic holographically reconstructed images with high gloss effects under continuous illumination. It achieves computational efficiency than existing algorithms, without increasing the number of polygons. Simulations and experimental results have verified the effectiveness of our proposed method. © 2024 SPIE. All rights reserved.

关键词： Rendering (computer graphics)

来源：评论

学校读者我要写书评

暂无评论

Prompt-Free Few-Shot Learning with ELECTRA for Acceptability Judgment 12th

Prompt-Free Few-Shot Learning with ELECTRA for Acceptabili...

引用

12th National CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2023

作者： Li, Linqin Li, Zicheng Chen, Ying Li, Shoushan Zhou, Guodong School of Computer Science and Technology Soochow University Jiangsu Suzhou215006 China College of Information and Electrical Engineering China Agricultural University Beijing Beijing100083 China

ISBN: (纸本)9783031446955

Few-shot learning remains a great challenge for the task of acceptability judgment that identifies whether a sentence is acceptable or unacceptable. In this paper, we propose a prompt-free learning approach, namely PF-ELECTRA, to few-shot acceptability judgment. First, we leverage a pre-trained token replaced detection model, ELECTRA, as our basic few-shot learner to deal with the challenge of data distribution difference. Second, we design a prompt-free few-shot learning strategy that uses both the maximal unacceptability score for a single token and the overall unacceptability score for the whole sentence to judge the acceptability. Empirical studies validate the effectiveness of PF-ELECTRA on challenging few-shot acceptability judgment. To the best of our knowledge, it is the first work that improves the performance of few-shot acceptability judgment based on standard fine-tuning. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Artificial intelligence

来源：评论

学校读者我要写书评

暂无评论

InvNets: A Novel Approach for Parkinson Disease Detection Using Involution Neural Networks 2

InvNets: A Novel Approach for Parkinson Disease Detection Us...

引用

2nd International Conference Computational and Characterization Techniques in engineering and sciences, IC3TES 2024

作者： Ayalew, Aleka Melese Salau, Ayodeji Olalekan Asnaked, Nigus Wereta Bezabh, Yohannes Agegnehu Gupta, Shashi Kant Israr, Mohammad Abuhayi, Biniyam Mulugeta Braide, Sepiribo Lucky University of Gondar Department of Information Technology Gondar Ethiopia Afe Babalola University Department of Electrical/Electronics and Computer Engineering Ado-Ekiti Nigeria Debark University Department of Information Technology Debark Ethiopia Eudoxia Research University Department of Computer Science and Engineering United States Maryam Abacha American University of Nigeria Kano State Hotoro GRA Nigeria Rivers State University Department of Electrical and Electronics Engineering Port Harcourt5080 Nigeria

ISBN: (纸本)9798350364699

Parkinson's disease (PD) is a progressive neurological ailment that requires early discovery for effective treatment. Traditional PD detection methods use convolutional neural networks (CNNs), which can be computationally intensive due to their huge number of parameters. This study proposes a novel approach for PD classification utilizing Involutional Neural Networks (InvNets), which reduce the parameter-intensive nature of CNNs. The involution kernel differs from the convolution kernel in that it is both location-specific and channel-agnostic. This location-specific operation enhances the network's ability to acquire detailed elements in medical images by adapting to diverse visual patterns across spatial regions. In that regard, magnetic resonance imaging (MRI) has become an important modality for PD detection. Furthermore, image preprocessing, data augmentation techniques, and the use of Block-Matching and 4D Filtering (BM4D) for noise filtering improve the robustness and accuracy of our proposed InvNets models. The proposed model results showed that InvNets outperformed traditional CNNs in classification tasks. The InvNets architecture has shown outstanding performance, achieving a training accuracy of 99.8%, validation accuracy of 98.5%, testing accuracy of 99.1%, precision 99%, recall 99%, and F1-score 99%. The suggested InvNets model is extremely effective for medical image processing, especially in situations with limited computer resources, as seen by improved accuracy and a lower parameter count. As a result, the study found that InvNets provided consistent and accurate features for breast cancer detection and quicker diagnosis. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of ChatGPT, DeepSeek, and Gemini for Automated Code Generation

Comparative Analysis of ChatGPT, DeepSeek, and Gemini for Au...

引用

International Conference on engineering of Modern Electric Systems (EMES)

作者： Khuram Shahzad Sohail Iqbal School of Electrical Engineering & Computer Science National University of Sciences and Technology (NUST) Islamabad Pakistan

ISBN: (数字)9798331525774

ISBN: (纸本)9798331525781

Code generation using artificial intelligence (AI) has revolutionized software development, providing automated coding solutions. This study conducts a systematic comparative analysis of three leading large language models (LLMs) such as ChatGPT (O1), DeepSeek (R1) and Gemini (2.0 Flash thinking), for Python code generation, evaluating their performance in correctness, code quality, and computational efficiency. Using a curated dataset of Codeforces programming problems that span various difficulty levels $(\mathbf{8 0 0 - 2 0 0 0}$ complexity), the research employs a rigorous evaluation framework that integrates online judge validation, static code analysis, and runtime profiling. The experimental results reveal that DeepSeek achieves comparatively higher correctness by consistently producing accepted solutions in fewer attempts, although its reasoning time increases with problem complexity. Gemini, on the other hand, is remarkably fast, delivering results in a fraction of the time, but its correctness deteriorates on more complex tasks. ChatGPT offers balanced performance with intermediate correctness and efficiency; however, it sometimes exhibits lower code quality. Overall, our findings underscore the inherent trade-offs between efficiency, accuracy, and quality in AI-generated code. The study provides actionable insights for developers, emphasizing the need to align model selection with project requirements, and contributes a replicable framework for future evaluations of AI code generation tools.

关键词： Hands Codes Systematics Runtime Large language models Chatbots Cognition Complexity theory Software development management Python

来源：评论

学校读者我要写书评

暂无评论

Two-timescale Optimization for E2E Network Slicing-aided Cloud-edge Collaborative Networks

引用

IEEE Transactions on Vehicular technology 2025年第6期74卷 9777-9789页

作者： Wang, Yunfeng Zhao, Liqiang Chu, Xiaoli Song, Shenghui Deng, Yansha Nallanathan, Arumugam Zhou, Guorong Xidian University State Key Laboratory of Integrated Services Networks Xi'an710071 China Xidian University Guangzhou Institute of Technology Guangzhou510100 China University of Sheffield Department of Electronic and Electrical Engineering SheffieldS10 2TN United Kingdom Department of Electronic and Computer Engineering Hong Kong University of Science and Technology Hong Kong King's College London Department of Engineering LondonWC2R 2LS United Kingdom Queen Mary University of London School of Electronic Engineering and Computer Science LondonE1 4NS United Kingdom Kyung Hee University Department of Electronic Engineering Gyeonggi-do Yongin-si17104 Korea Republic of

To leverage the synergy between cloud computing (CC) and edge computing (EC) to support various services while reducing the CC/EC switching overhead, the two-timescale utility maximization problem for an end-to-end network slicing-aided cloud-edge collaborative network (E2E-CECN) is formulated covering both the high throughput and low delay service requirements. To solve the utility maximization problem while dynamically adjusting the weights of the E2E-CECN utility to accommodate the variation of users' service requests, we proposed a reward comparison double deep Q network algorithm to optimize the large timescale joint virtual base station activation and CC-EC scheduling, and a reward comparison deep deterministic policy gradient algorithm to optimize the small timescale allocation of backhaul link capacity (BLC), CC/EC capability and transmission power. Numerical results show that both the high-throughput and low-delay service requirements can be satisfied simultaneously under the reasonable BLC usage and power consumption. © 1967-2012 IEEE.

关键词： Cloud platforms

来源：评论

学校读者我要写书评

暂无评论

6-D Movable Antenna Enhanced Interference Mitigation for Cellular-Connected UAV Communications

引用

IEEE Wireless Communications Letters 2025年第6期14卷 1618-1622页

作者： Ren, Tianshi Zhang, Xianchao Zhu, Lipeng Ma, Wenyan Gao, Xiaozheng Zhang, Rui Beijing Institute of Technology School of Information and Electronics Beijing100081 China National University of Singapore Department of Electrical and Computer Engineering Queenstown 117583 Singapore Jiaxing University Provincial Key Laboratory of Multimodal Perceiving and Intelligent Systems Engineering Research Center of Intelligent Human Health Situation Awareness of Zhejiang Province Jiaxing314001 China School of Science and Engineering Guangdong Shenzhen518172 China

Cellular-connected unmanned aerial vehicle (UAV) communications is an enabling technology to transmit control signaling or payload data for UAVs through cellular networks. Due to the line-of-sight dominant air-to-ground channels, efficient interference mitigation is crucial to UAV communications, while the conventional fixed-position antenna (FPA) arrays have limited degrees of freedom to suppress the interference between the UAV and its non-associated co-channel base stations (BSs). To address this challenge, we propose in this letter a new approach by utilizing the six-dimensional movable antenna (6DMA) arrays to enhance the interference mitigation for the UAV. Specifically, we propose an efficient block coordinate descent algorithm to iteratively optimize the antenna position vector, array rotation vector, receive beamforming vector, and associated BS of the UAV to maximize its signal-to-interference-plus-noise ratio. Numerical results show that the proposed 6DMA enhanced cellular-connected UAV communication can significantly outperform that with the traditional FPA arrays and other benchmark schemes in terms of interference mitigation. © 2012 IEEE.

关键词： Aircraft communication

来源：评论

学校读者我要写书评

暂无评论

Insect Identification in Pulsed Lidar Images Using Changepoint Detection Algorithms

Insect Identification in Pulsed Lidar Images Using Changepoi...

引用

2023 Annual Intermountain engineering, technology and Computing, IETC 2023

作者： Sweeney, Nathaniel Xu, Caroline Shaw, Joseph A. Hocking, Toby D. Whitaker, Bradley M. Montana State University Electrical and Computer Engineering Department United States University of Michigan Electrical Engineering and Computer Science Department United States Montana State University Optical Technology Center United States School of Informatics Computing and Cyber Systems Northern Arizona University United States

ISBN: (纸本)9798350335903

Noninvasive entomological insect monitoring often utilizes a variety of tools such as LiDAR to gather information without interfering with the insects in their habitat. These collection methods often result in large amounts of data that can be te-dious and lengthy to interpret and analyze. Machine learning has been previously used in the past in order to analyze Li-DAR images to detect insects, but often suffers from pitfalls such as long training times and large computational power requirements. In an attempt to offer an alternative that takes little to no training on the data and much less computational power, this paper looks at the use of changepoint detection algorithms to analyze LiDAR images containing insects. By analyzing the rows or columns of a LiDAR image, the algorithms should be able to detect abrupt changes in the image that would represent the insects. While not as accurate, the changepoint detection algorithms give comparable results to a machine learning algorithm tested on the same dataset without the need for supervised training. © 2023 IEEE.

关键词： Anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Robust sparse regression with non-isotropic designs 24

Robust sparse regression with non-isotropic designs

引用

Proceedings of the 38th International Conference on Neural Information Processing Systems

作者： Chih-Hung Liu Gleb Novikov Department of Electrical Engineering National Taiwan University Lucerne School of Computer Science and Information Technology

ISBN: (纸本)9798331314385

We develop a technique to design efficiently computable estimators for sparse linear regression in the simultaneous presence of two adversaries: oblivious and adaptive. Consider the model y* = X* β* + η where X* is an n × d random design matrix, β* ∈ ℝd is a k-sparse vector, and the noise η is independent of X* and chosen by the oblivious adversary. Apart from the independence of X*, we only require a small fraction entries of η to have magnitude at most 1. The adaptive adversary is allowed to arbitrarily corrupt an ε-fraction of the samples (X*1, y*1),..., (X*n, y*n). Given the e-corrupted samples (X1, y1),..., (Xn, yn), the goal is to estimate β*. We assume that the rows of X* are iid samples from some d-dimensional distribution Ɗ with zero mean and (unknown) covariance matrix σ with bounded condition *** design several robust algorithms that outperform the state of the art even in the special case of Gaussian noise η ~ N(0, 1)n. In particular, we provide a polynomial-time algorithm that with high probability recovers β* up to error O(√ε) as long as n ≥ Õ(k2/ε), only assuming some bounds on the third and the fourth moments of Ɗ. In addition, prior to this work, even in the special case of Gaussian design Ɗ = N(0, Σ) and noise η ~ N(0, 1), no polynomial time algorithm was known to achieve error o(√ε) in the sparse setting n < d2. We show that under some assumptions on the fourth and the eighth moments of Ɗ, there is a polynomial-time algorithm that achieves error o(√ε) as long as n ≥ Õ(k4/ε3). For Gaussian distribution Ɗ = N(0, σ), this algorithm achieves error O(ε3/4). Moreover, our algorithm achieves error o(√ε) for all log-concave distributions if √ ≤ 1/polylog(d).Our algorithms are based on the filtering of the covariates that uses sum-of-squares relaxations, and weighted Huber loss minimization with ℓ1 regularizer. We provide a novel analysis of weighted penalized Huber loss that is suitable for heavy-tailed designs in the presence of two adversaries. Fu

关键词：

来源：评论

学校读者我要写书评

暂无评论

Comparison between Single-Input and Multi-Input Classification with the Application of Canny Feature Extraction and Classification Algorithms on the Toraja Buffalo Dataset 19

Comparison between Single-Input and Multi-Input Classificati...

引用

19th International Conference on Ubiquitous Information Management and Communication, IMCOM 2025

作者： Manga, Abdul Rachman Nanda, Assyahrin Handayani, Anik Nur Herwanto, Heru Wahyu Asmara, Rosa Andrie Lantara, Dirgahayu Department of Computer Science Universitas Muslim Indonesia Makassar Indonesia Department of Electrical Engineering Universitas Negeri Malang Malang Indonesia Information Technology Department State Polytechnic of Malang Malang Indonesia Faculty of Industrial Engineering Universitas Muslim Indonesia Makassar Indonesia

ISBN: (纸本)9798331507817

Data classification plays a crucial role in artificial intelligence, particularly in enhancing model accuracy. This study focuses on classifying Toraja buffalo, a livestock breed with significant cultural importance in South Sulawesi, Indonesia. While the Single Input Approach is commonly used for classification, it often fails to capture all the necessary attributes to effectively distinguish between racial traits. Therefore, this research aims to evaluate the effectiveness of a multi-input approach, which integrates multiple data inputs to improve classification performance compared to the Single Input method. We employed four classification techniques: K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Naïve Bayes, and Decision Tree, using both Single Input and Multi Input configurations. Model performance was assessed through Precision, Recall, F1 Score, and Accuracy metrics. The findings indicate that the Multi Input approach consistently outperformed the Single Input method. Notably, KNN achieved its best performance with Multi Input, recording an F1 Score of 0.7263 and an Accuracy of 0.7333, significantly surpassing the results obtained from Single Input. Similarly, SVM also demonstrated substantial performance enhancements with Multi Input. Overall, the study highlights the importance of incorporating a wider array of informative data to enhance the model's capability in accurately classifying specific categories, with KNN showing the most pronounced improvements. © 2025 IEEE.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Multispectral Registration Algorithm Based on Fusion of Infrared and Color Images 16

Multispectral Registration Algorithm Based on Fusion of Infr...

引用

16th International Conference on Cyber-Enabled Distributed Computing and Knowledge Discover, CyberC 2024

作者： Wan, Teng Du, Shaoyi Zhang, Qiang Qi, Ying Li, Ce Northwest Normal University The College of Computer Science and Engineering Lanzhou China Xi'an Jiaotong University The Institute of Artificial Intelligence and Robotics The College of Artificial Intelligence Xi'an China Lanzhou University of Technology The College of Electrical and Information Engineering Lanzhou China

ISBN: (纸本)9798331506896

Multispectral data fusion is a fundamental research in the fields of artificial intelligence and computer vision, which provides technical support and theoretical basis for remote sensing, autonomous driving, and medical imaging. In this paper, we present a robust registration algorithm to align RGB-D scans. In colored point clouds, limited by the principle of point cloud collection and generation, noise, and outliers are common and may cause registration failure. Therefore, the core of our method is to apply color information and kernel function to remove the influence of noise and outliers on the alignment accuracy. In the beginning, for the first time, we innovatively use the Cauchy kernel as the energy function of the point set registration to assign different weights to inliers, noise, and outliers to maximize their differentiation. Combined with the maximization optimization strategy, noise and outliers are removed to achieve robust registration. Secondly, since RGB-D data contain both shape and color information, the trimmed bidirectional color distance is applied to reduce the probability of noise and outliers being selected as the matching point pairs when establishing the correspondences. Moreover, the overall formulation is solved by the iterative optimization method. Experiments demonstrate that our algorithm is more robust and accurate than previous point set registration algorithms, especially when aligning point sets containing noise and outliers. © 2024 IEEE.

关键词： Image registration

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：