检索结果-内蒙古大学图书馆

2024 Conference on Empirical Methods in Natural Language Processing, EMNLP 2024

作者： Lu, Xiaotian Li, Jiyi Takeuchi, Koh Kashima, Hisashi Department of Intelligence Science and Technology Graduate School of Informatics Kyoto University Kyoto Japan Department of Computer Science and Engineering University of Yamanashi Kofu Japan

ISBN: (纸本)9798891761681

Question answering (QA) tasks have been extensively studied in the field of natural language processing (NLP). Answers to open-ended questions are highly diverse and difficult to quantify, and cannot be simply evaluated as correct or incorrect, unlike close-ended questions with definitive answers. While large language models (LLMs) have demonstrated strong capabilities across various tasks, they exhibit relatively weaker performance in evaluating answers to open-ended questions. In this study, we propose a method that leverages LLMs and the analytic hierarchy process (AHP) to assess answers to open-ended questions. We utilized LLMs to generate multiple evaluation criteria for a question. Subsequently, answers were subjected to pairwise comparisons under each criterion with LLMs, and scores for each answer were calculated in the AHP. We conducted experiments on four datasets using both ChatGPT-3.5-turbo and GPT-4. Our results indicate that our approach more closely aligns with human judgment compared to the four baselines. Additionally, we explored the impact of the number of criteria, variations in models, and differences in datasets on the results. © 2024 Association for Computational Linguistics.

关键词： Computational linguistics

来源：评论

学校读者我要写书评

暂无评论

An Enhanced Optimization of Automated Detection of Cardiac Abnormalities Using Deep Learning

An Enhanced Optimization of Automated Detection of Cardiac A...

引用

2024 International Conference on Optimization Computing and Wireless Communication, ICOCWC 2024

作者： Ojha, Ananta Yadav, Dhyan Chandra Goyal, Manish Kumar Department of Computer Science and Information Technology Bangalore India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India Vivekananda Global University Department of Computer Science & Application Jaipur India

ISBN: (纸本)9798350383348

Deep gaining knowledge of is a place of artificial intelligence that is becoming increasingly famous inside the scientific area. This paper provides an improved optimization of computerized detection of cardiac abnormalities the use of deep learning. particularly, the authors recommend using a convolutional neural community (CNN) to stumble on abnormalities from ECG records. They use an ensemble of models to further improve accuracy and reduce false superb quotes. moreover, they apply transfer learning techniques to higher generalize the mastering from the EEG facts. The authors take a look at their optimized set of rules on two datasets of ECG recordings and file an normal accuracy of 88.9%. This demonstrates the potential for deep getting to know techniques to end up an increasing number of reliable and sturdy for detecting cardiac abnormalities. The authors also talk the feasible directions of future studies and the potentials of deep learning for clinical safety and diagnostics in phrases of fee and efficiency. © 2024 IEEE.

关键词： abnormalities accuracy diagnostics specifically techniques

来源：评论

学校读者我要写书评

暂无评论

A Few-shot Classification Method with Coordinate Attention and Cross Attention 6

A Few-shot Classification Method with Coordinate Attention a...

引用

6th International Conference on Artificial Intelligence and Big Data, ICAIBD 2023

作者： Shu, Kai Hefei University of Technology School of Computer Science and Information Engineering Hefei China

ISBN: (纸本)9781665491259

Deep learning tries to learn embedding functions from a large amount of data. A lot of work shows that deep learning has surpassed humans in areas such as image classification and image segmentation. Few-shot learning focuses on learning from unknown classes with limited labels, and the data scarcity of few-shot learning is destined to be a very challenging task. Our work studies the role of the attention mechanism in few-shot learning. In the pre-training process of the two-stage few-shot learning, this study introduces the coordinate attention mechanism to improve the ability of the model to focus on key areas. In the subsequent meta-training process, this study applies the cross-attention module to the support set embedding and the query set embedding, which generates cross-attention maps for each pair of class features and query sample features, alleviating the problem that the Transformer focuses more on the support set features during the meta-training process. Our work has significantly improved the accuracy rate based on the original works and is close to the existing state-of-the-art performance on the general few-shot classification dataset miniImageNet and tieredImageNet. © 2023 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

A Sustainable WSN System with Heuristic Schemes in IIoT

引用

computers, Materials & Continua 2022年第9期72卷 4215-4231页

作者： Wenjun Li Siyang Zhang Guangwei Wu Aldosary Saad Amr Tolba Gwang-jun Kim School of Computer and Communication Engineering Changsha University of Science and TechnologyChangsha410114China College of Computer and Information Engineering Central South University of Forestry and TechnologyChangsha410114China Computer Science Department Community CollegeKing Saud UniversityRiyadh11437Saudi Arabia Mathematics and Computer Science Department Faculty of ScienceMenoufia UniversityEgypt Department of Computer Engineering Chonnam National UniversityGwangju61186Korea

Recently, the development of Industrial Internet of Things hastaken the advantage of 5G network to be more powerful and more ***, the upgrading of 5G network will cause a variety of issues increase,one of them is the increased cost of coverage. In this paper, we proposea sustainable wireless sensor networks system, which avoids the problemsbrought by 5G network system to some extent. In this system, deployingrelays and selecting routing are for the sake of communication and *** main aim is to minimize the total energy-cost of communication underthe precondition, where each terminal with low-power should be charged byat least one relay. Furthermore, from the perspective of graph theory, weextract a combinatorial optimization problem from this system. After that,as to four different cases, there are corresponding different versions of theproblem. We give the proofs of computational complexity for these problems,and two heuristic algorithms for one of them are proposed. Finally, theextensive experiments compare and demonstrate the performances of thesetwo algorithms.

关键词： Industrial Internet of Things sustainable wireless sensor network system combinatorial optimization problem heuristic algorithms

来源：评论

学校读者我要写书评

暂无评论

Image Inpainting Based on Dual Dynamic Predictive Filtering and Fast Fourier Convolution

Image Inpainting Based on Dual Dynamic Predictive Filtering ...

引用

2023 International Conference on New Trends in Computational Intelligence, NTCI 2023

作者： Xie, Zhengrong Zhang, Zhaoyuan Zhang, Jianming School of Computer and Communication Engineering Changsha University of Science and Technology Changsha China

ISBN: (纸本)9798350380859

Despite outstanding performances of the existing image inpainting methods in restoring corrupted images, they often generate images with artifacts or missing details. Predictive filtering is a widely used image restoration technique which can preserve local structures and avoid artifacts, but it cannot fill large missing areas. Convolutional neural networks are suitable for large missing areas, but it is difficult to restore fine details. Therefore, we treat image inpainting as a filtering task, and propose a novel image inpainting method consisting of two networks, namely the Attention-Enhanced Kernel Predictive Network (AE-KPNet) and the Deep Generative Network (DGNet). The two parallel networks are linked, and DGNet provides features to AE-KPNet, while AE-KPNet trains dual dynamic predictive filtering kernels for DGNet. To address the issue of large missing areas, we introduce fast Fourier convolution into DGNet, and the operation in the frequency domain expands the spatial receptive field. We evaluate our model on three public datasets, CelebA, Places2, and Paris StreetView. Through quantitative and qualitative experiments, the results show that our method achieves the performance of state-of-the-art methods. © 2023 IEEE.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Multilevel Discrete Cosine Transform and CNN-Based Image Steganography 3

Multilevel Discrete Cosine Transform and CNN-Based Image Ste...

引用

3rd International Conference on Intelligent Systems, Advanced Computing, and Communication, ISACC 2025

作者： Subburaj, Brindha Sai, Aditya Shantanu, Shantanu Won, Daehan Vellore Institute of Technology School of Computer Science and Engineering Chennai India Binghamton University Department of Systems Science and Industrial Engineering United States

ISBN: (纸本)9798331523893

This paper investigates the growing role of steganography in cybersecurity and presents a hybrid implementation called Multi-level Discrete Cosine Convolution (MDCC) that applies a Multi-level Discrete Cosine Transform (MDCT) and a Convolutional Neural Network (CNN). Steganography provides some advantages compared to cryptography, mainly its invisibility, but there are weaknesses in information capacity and resistance to attacks in traditional techniques. The paper suggests solving these difficulties with a new steganography method using deep learning. The proposed MDCC method provides an opportunity to encrypt large amounts of data in realistic images that are beyond human visual perception. The system is trained on the ImageNet dataset by using a loss function to balance the quality of the cover image and the information it conceals within it. The MDCC model proposed yields accurate results for all metrics since PSNR, SSIM, and bits per pixel come out to be 80.13dB, 0.9918, and 12.5 bpp, respectively. © 2025 IEEE.

关键词： Discrete cosine transforms

来源：评论

学校读者我要写书评

暂无评论

The Enhanced Analysis of FACT-Based Framework for Traffic Prediction in V2V Networks

The Enhanced Analysis of FACT-Based Framework for Traffic Pr...

引用

2024 International Conference on Optimization Computing and Wireless Communication, ICOCWC 2024

作者： Inzimam Aadil, Mir Agarwal, Ankita Vivekananda Global University Department of Computer Science & Application Jaipur India Department of Computer Science and Information Technology Bangalore India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India

ISBN: (纸本)9798350383348

The vehicular-to-car (V2V) communique networks have enabled a new generation of urban transportation phase. Vehicular conversation technology can offer scalability, dependability, and accuracy in various vehicle offerings. Even though a few modern studies exist on predicting traffic behavior in vehicular networks, only some are primarily based on an analytical method., this paper presents a better analytical method primarily based on the Fuzzy Adaptive Cluster-based total model (fact) for visitor prediction in V2V networks. The truth-based total version includes an extensive fuzzy-common sense-based device to degree the continued conduct and site visitors for a selected target node and then extrapolate the conduct and site visitors for the upcoming frames. Moreover, the model also contains a k-method algorithm to prepare the visitor's statistics into clusters for analysis. The effects of the simulation indicate that the proposed truth-based version is capable of addressing greater precisely and accurately the site visitors' prediction in dynamic site visitor's conditions, attaining higher performance metrics in terms of prediction accuracy, latency, and jitter as compared to techniques, like k-Nearest Neighbor (KNN), assist Vector Machines (SVM), and Multi-Layer Perceptron (MLP). The model has primary importance for its capability as it should be degree and expects V2V site visitors in V2V communications. © 2024 IEEE.

关键词： communique networks dynamic predicting transportation

来源：评论

学校读者我要写书评

暂无评论

Improving Efficiency and Accuracy in Clinical Diagnostic Decision Making with Deep Learning

Improving Efficiency and Accuracy in Clinical Diagnostic Dec...

引用

2024 International Conference on Optimization Computing and Wireless Communication, ICOCWC 2024

作者： Kharbas, Vikas Kumar Gobi, N. Kumar, Dinesh Vivekananda Global University Department of Computer Science & Application Jaipur India Department of Computer Science and Information Technology Bangalore India Maharishi School of Engineering & Technology Maharishi University of Information Technology Uttar Pradesh India

ISBN: (纸本)9798350383348

Deep learning is a modern form of synthetic intelligence (AI) based on large quantities of statistics and algorithms that allow computer systems to approximate features and solutions, ensuing in more excellent, efficient, and accurate selections. In the scientific field, deep gaining knowledge is increasingly used to help clinicians with diagnostic decision-making, decreasing the value and attempt associated with diagnostic approaches. This paper discusses the potential of deep learning to enhance the performance and accuracy of medical diagnostic selection-making. The paper provides an overview of the cutting-edge nation of the field, with a particular awareness of picture processing and its application to improve accuracy and decrease the price associated with scientific diagnoses. It also provides a few capacity challenges and pitfalls and outlines the issues essential to ensure a hit implementation and successful results. Eventually, it offers a few recommendations for similar studies to increase deep getting to know algorithms further. Deep gaining knowledge has been broadly explored in recent years as a powerful tool for enhancing the accuracy and efficiency of clinical diagnostic decision-making. Deep studying models can investigate patterns in medical datasets, deliberating the traits of affected person information and ailment signs to generate extra specific and reliable diagnoses. By incorporating extra facts such as transcripts from affected person-clinician interactions and medical imaging, those fashions may be trained to assist clinicians in identifying illnesses early and provide well-timed and correct analysis. Moreover, deep learning allows for incorporating case-primarily based reasoning and presents the capacity to combine scientific knowledge in real-time. By having extra sturdy and accurate models, clinicians can reduce the time required to diagnose a condition and make treatment choices. Deep gaining knowledge also lets in for the integration o

关键词： disorder enhance healthcare value

来源：评论

学校读者我要写书评

暂无评论

Comparisons of Filtered Back-projection and Time Reversal Algorithms in Photoacoustic Tomography 2

Comparisons of Filtered Back-projection and Time Reversal Al...

引用

2nd Conference on Biomedical Photonics and Cross-Fusion, BPC 2023

作者： Wei, Jin Shen, Kang Tian, Chao School of Engineering Science University of Science and Technology of China Anhui Hefei230026 China Institute of Artificial Intelligence Hefei Comprehensive National Science Center Anhui Hefei230088 China

ISBN: (数字)9781510667556

ISBN: (纸本)9781510667549

In photoacoustic tomography (PAT), image reconstruction has a fundamental impact on image quality and imaging speed. Among various reconstruction algorithms, the analytical filtered back-projection (FBP) and the numerical time reversal (TR) algorithms are two commonly used image reconstruction techniques in PAT. However, so far, no comprehensive studies are reported on the comparisons of FBP and TR algorithms. In this work, we compare these two algorithms from the perspectives of computational efficiency, robustness to non-ideal detection surfaces, and applicability to heterogeneous media. The results show that: 1) In terms of computational efficiency, FBP is typically faster than TR due to its flexibility in the selection of the reconstruction region. 2) For non-ideal detection surfaces-based reconstruction, FBP can provide more accurate amplitude information for the limited-view reconstruction and can produce fewer image artifacts for the sparse-view reconstruction. 3) For acoustically heterogeneous media, the TR algorithm can incorporate acoustic properties and thus can yield high-quality images, while FBP fails in this case. This study can help researchers gain a deeper understanding of the FBP and TR algorithms and is expected to provide a guide for the reasonable selection of PAT image reconstruction algorithms. © 2023 SPIE.

关键词： Image reconstruction

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Pricing Method with LoRa Technique in Smart Parking Systems Along with the Edge-Cloud Based Computing 1

A Dynamic Pricing Method with LoRa Technique in Smart Parkin...

引用

1st International Conference on Innovative Sustainable Technologies for Energy, Mechatronics and Smart Systems, ISTEMS 2024

作者： Singh, Baldev Kumar, Ritesh Neetha, S.S. Vivekananda Global University Department of Computer Science and Engineering Jaipur India Maharishi School of Engineeing and Technology Maharishi University of Information Technology Uttar Pradesh India Department of Computer Science and Information Technology Bangalore India

ISBN: (纸本)9798350384246

The increasing quantity of automobiles on the road these days has resulted in chronic traffic jams and difficulties with urban traffic control. Increased traffic bottlenecks are mostly caused by unlicensed and unregulated parking. The management of present parking spots is ineffective, and traffic-related problems are made worse by a lack adequate real-time advice and general ignorance. The constrained growth of highways and other transportation and related infrastructure highlights the urgent need for a clever, flexible, and effective parking solution. We suggest a complex multi-layer design for intelligent parking facilities to overcome these issues. This system uses cutting-edge extended, low-power wireless communications, powerful edge and cloud computing processing capacity, and sensor nodes with different parameters for parking spots. By dynamically managing large areas and providing drivers with real-time information on parking spaces and related services via continuous vehicle tracking, our proposed system seeks to transform parking management. We also provide a dynamic pricing strategy that aims to maximise parking authority income while guaranteeing drivers have the best possible availability of parking spaces. By addressing the problems caused by careless parking and improving general traffic control, this all-encompassing strategy aims to improve the effectiveness and smooth operation of the city's transport network. © 2024 IEEE.

关键词： Traffic congestion

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：