检索结果-内蒙古大学图书馆

Adaptive Genghis Khan Shark based convolutional denoising autoencoder for quality enhancement of satellite images

International Journal of Information Technology (Singapore) 2025年 1-14页

作者： Rajalakshmi, M. Subramanian, R. S. Sankara Tamilselvan, S. Bhuvanesh, A. Department of Computing Technologies School of Computing SRM Institute of Science and Technology Tamil Nadu Chennai India Department of Mathematics PSG Institute of Technology and Applied Research Tamil Nadu Coimbatore India Department of Electronics and Communication Engineering Sri Eshwar College of Engineering Tamil Nadu Coimbatore India Department of Electrical and Electronics Engineering PSN College of Engineering and Technology Tamil Nadu Tirunelveli India

Satellite image processing often enhances feature perception and visibility for exact feature identification. Additionally, the change improves remote sensing data quality and clarity. Due to their long distances, these photos often have weather-related noise and distortion that must be fixed. This paper addresses these challenges with a Deep Learning (DL) satellite image denoising technique. An Adaptive Genghis Khan Shark Algorithm (AGKSA) and a Convolutional Denoised Autoencoder are used to improve satellite images in this study. The CDAE settings are optimized using the AGKSA while keeping important image attributes. AGKSA is used in this hybrid model to optimize CDAE performance under various noise conditions. Experimental results showed that the proposed model outperforms traditional image denoising approaches in satellite imagery feature extraction and visual quality. In very precise image processing applications like environmental monitoring and land use classification, the CDAE-AGKSA works well. © Bharati Vidyapeeth's Institute of Computer Applications and Management 2025.

关键词： Adaptive Genghis Khan algorithm Convolutional denoised autoencoder Satellite image enhancement Satellite image processing

来源：评论

学校读者我要写书评

暂无评论

Deep Learning and IoT-Enabled Accident Detection and Reporting for Smart Cities Domain

引用

KSII Transactions on Internet and Information Systems 2025年第4期19卷 1140-1166页

作者： Javed, Rawal Mazhar, Tehseen Aoun, Muhammad Shahzad, Tariq Al-AlShaikh, Halah A. Ghadi, Yazeed Yasin Saudagar, Abdul Khader Jilani Khan, Muhammad Amir School of Automation Central South University Changsha China School of Computer Science National College of Business Administration and Economics Lahore54000 Pakistan Department of Computer Science and Information Technology School Education Department Government of Punjab Layyah31200 Pakistan Department of CS IT Ghazi University Punjab Dera Ghazi Khan Pakistan School of Electrical Engineering Dept. of Electrical and Electronic Eng. Science University of Johannesburg Johannesburg2006 South Africa Riyadh11432 Saudi Arabia Department of Computer Science and Software Engineering Al Ain University Abu Dhabi12555 United Arab Emirates School of Computing Sciences College of Computing Informatics and Mathematics University Technology MARA Selangor Shah Alam40450 Malaysia

Even though numerous ICT-based solutions have been put forth for the detection of accidents and rescue missions, they often suffer from compatibility issues with different vehicles and are accompanied by high costs. This paper introduces a cutting-edge intelligent system designed for accident detection and rescue, mirroring human cognitive functions through integrating the Internet of Things (IoT) and artificial intelligence. A specialized kit for IoT has been created to identify accidents and gather comprehensive incident-related data, such as pressure, position, gravitational force, and speed. This information is then transmitted to the cloud for processing. A deep learning model is implemented in the cloud environment to verify the IoT module's output and activate the rescue mechanism when an accident is detected. The DL module, employing ensemble transfer learning with dynamic weights, aims to minimize false detection rates. Because there isn't an appropriate dataset available, a personalized dataset is generated from diverse online videos. To validate the proposed method, a comparative analysis is conducted on Yolo V4. Experimental results indicate that Yolo V4 outperforms, achieving training, validation, and test accuracies of 98.2% each. To assess the system's real-world applicability, it undergoes validation using a robot car model. Copyright © 2025 KSII.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Grad-CAM Guided Framework for Quality-Preserving Channel Pruning in GANs

Grad-CAM Guided Framework for Quality-Preserving Channel Pru...

引用

National Conference on Communications (NCC)

作者： Pallav Dwivedi Parimala Kancharla School of Computing and Electrical Engineering IIT Mandi Himachal Pradesh India

ISBN: (数字)9798331536015

ISBN: (纸本)9798331536022

The remarkable performance of Generative Adversarial Networks (GANs) in tasks such as picture synthesis and editing is well known. However, especially in contexts with limited resources, their massive structures and high parameter counts pose serious computing hurdles. Here, we provide a Grad-CAM Guided technique to prune StyleGan, a unique method that computes channel-wise significance scores by utilizing Grad-CAM, a class-discriminative localization technique. This allows us to efficiently trim channels without sacrificing visual quality. Our methodology prioritizes perceptual quality over standard content-based or diversity-driven pruning techniques, striking a compromise between computational efficiency and output fidelity. Our approach uses Grad-CAM maps to emphasizes important regions in the generated sample and helps in computing the channel sensitivity better and incorporates the results into a channel-wise pruning plan. StyleGAN models using popular datasets such as FFHQ-256, LSUN-HORSE, and LSUN-CHURCH have been used to validate the framework. Our method's robustness is highlighted by experimental results that show large parameter count reductions while maintaining competitive Fréchet Inception Distance (FID) scores. This work shows how Grad-CAM may be used to guide quality-aware pruning techniques and provides a fresh viewpoint on generative model optimisation by fusing interpretability with computational efficiency. This methodology has more applicability in real-world scenarios because it can be tailored to different GAN designs.

关键词： Location awareness Visualization Sensitivity Quantization (signal) Computational modeling Generative adversarial networks Robustness Computational efficiency Standards Optimization

来源：评论

学校读者我要写书评

暂无评论

Detecting Concept Shifts Under Different Levels of Self-awareness on Emotion Labeling 27th

Detecting Concept Shifts Under Different Levels of Self-awa...

引用

27th International Conference on Pattern Recognition, ICPR 2024

作者： Choi, HyoSeon Choi, Dahoon Kaongoen, Netiwit Kim, Byung Hyung Department of Electrical and Computer Engineering Incheon Korea Republic of Department of Computer Engineering Incheon Korea Republic of Department of Artificial Intelligence Inha University Incheon Korea Republic of School of Computing KAIST Daejeon Korea Republic of

ISBN: (纸本)9783031782008

Generalizing deep learning for all requires individual self-assessment. However, the quality of ground-truth labels depends on the annotators’ self-awareness. Real-world datasets inevitably experience the Concept Shift problem. Recent advances in Out-of-distribution (OOD) detection have received much attention due to its ability to alleviate distribution shift problems by distinguishing between anomalous and in-distribution(ID) data samples. Existing approaches underlie pre-trained ID models learned with class-balanced data. However, this assumption makes the methods incapable when the ID models are trained with inter- and intra-class variance depending on user characteristics, such as gender, culture, and genetics. We present an OOD detection framework. Our system builds a generalized ID model by extracting high-quality data from high-dimensional neural activities considering individuals’ cognitive and perceptional ability to evaluate self-assessments. The proposed system detects and removes abnormal pairs of data and labels to enhance model performance by considering the maximum softmax probability approach. Experimental results on public EEG datasets in emotion recognition demonstrate the superiority of our method despite the non-stationary nature of EEG signals. The codes are available at https://***/affctivai/coglier. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Emotion Recognition

来源：评论

学校读者我要写书评

暂无评论

Parameter Adaptation of Enhanced Ant Colony System for Water Quality Rules Classification

引用

International Journal of Advanced Computer Science and Applications 2025年第3期16卷 716-723页

作者： Nasir, Husna Jamal Abdul Munif, Mohd Mizan Ahmad, Muhammad Imran Chow, Tan Shie Ku-Mahamud, Ku Ruhana Abdullah, Abu Hassan Faculty of Electronic Engineering and Technology Universiti Malaysia Perlis Perlis Malaysia Universiti Malaysia Perlis Perlis Malaysia Faculty of Business Management and Information Technology Universiti Muhammadiyah Malaysia Perlis Malaysia School of Computing Universiti Utara Malaysia Sintok Kedah Malaysia Faculty of Electrical Engineering & Technology Universiti Malaysia Perlis Perlis Malaysia

Water quality monitoring in aquaculture involves classifying and analyzing the collected data to assess the water quality that is appropriate for breeding, rearing and harvesting aquatic organisms. Systematic data classification is essential when it comes to managing large amounts of data that are continuously sensed in real time and have various attributes in each instance of a sequence. Ant Colony System (ACS) has been employed in optimizing the data classification in smart aquaculture, where the majority of the research focuses on enhancing the classification procedure using predetermined parameters within a specified range. Nevertheless, this approach does not guarantee ideal performance. This paper enhances the ACS algorithm by introducing the Enhanced Ant Colony System-Rule Classification (EACS-RC) algorithm, which improves rule construction by integrating pheromone and heuristic values while incorporating advanced pheromone update techniques. The optimal parameter values to be used by the proposed algorithm are obtained from parameter adaptation experiments in which different values within the defined range were applied to obtain the optimal value for each parameter. Experiments were performed on the Kiribati water quality dataset and the results of the EACS-RC algorithm were evaluated against the AntMiner and AGI-AntMiner algorithms. Based on the results, the proposed algorithm outperforms the benchmark algorithms in classification accuracy and processing time. The output of this study can be adopted by the other ACS variants to achieve optimal performance for data classification in smart aquaculture. © (2025), (Science and Information Organization). All rights reserved.

关键词： Aquaculture

来源：评论

学校读者我要写书评

暂无评论

Use of deep geothermal energy to supply heat to a village in Scotland

引用

Interactions 2025年第1期246卷 1-18页

作者： Penman, David Muhammad-Sukki, Firdaus Abu-Bakar, Siti Hawa Bani, Nurul Aini Mas’ud, Abdullahi Abubakar School of Computing Engineering & the Built Environment Edinburgh Napier University EdinburghEH10 5DT United Kingdom Renewable Energy Research Laboratory Electrical Engineering Section British Malaysian Institute Universiti Kuala Lumpur Gombak53100 Malaysia Faculty of Artificial Intelligence Universiti Teknologi Malaysia Jalan Sultan Yahya Petra Kuala Lumpur54100 Malaysia Department of Electrical Engineering Jubail Industrial College Jubail Industrial City35718 Saudi Arabia

This paper explores the possibility of utilising geothermal energy in an existing district heat network in the Aberdeenshire village of Banchory to provide hot water through the creation of a hypothetical geothermal well. It considers a simulated deep geothermal single well (DGSW), commonly referred to as a coaxial system with a depth of 5000 m. The hypothetical well was created using the dimensions of existing oil wells which, once they have reached the end of their lifecycle, hold the potential to be repurposed for geothermal use. It was found that the thermal output of the well decreases over time due to the drop in local rock temperature which is a result of thermal extraction. Given this, the thermal output after a year of operation was calculated and found that, to directly supply the Banchory heat network, a volumetric flow rate of 4.8m3/hr was required. After a year of operation, the site's peak thermal production was 108.4 kW which equates to a production of more than 949.9 MWh over the first operational year. © The Author(s) 2025.

关键词： Geothermal wells

来源：评论

学校读者我要写书评

暂无评论

Using Large Language Models to Mitigate Human-Induced Bias in SMS Spam: An Empirical Approach 4

Using Large Language Models to Mitigate Human-Induced Bias i...

引用

4th IEEE International Conference on Artificial Intelligence in Cybersecurity, ICAIC 2025

作者： Fuhnwi, Gerard Shu Revelle, Matthew Whitaker, Bradley Izurieta, Clemente Montana State University Gianforte School of Computing BozemanMT United States Montana State University Electrical & Computer Engineering BozemanMT United States Montana State University Pacific Northwest National Laboratory Idaho National Laboratory Gianforte School of Computing BozemanMT United States

ISBN: (纸本)9798331518882

Short Message Service (SMS) is a widely used text messaging feature on both basic and smartphones. SMS spam detection is a crucial task. Traditional machine learning approaches often struggle in this domain due to their reliance on manually crafted features, such as keyword detection, which can result in overly simplistic patterns and misclassification of more complex messages. With this shortcoming, these models can amplify human-induced biases if the training data contains inconsistent labeling or subjective interpretations, leading to unfair treatment of specific keywords or contexts. Conversely, advanced LLMs present effective approaches to addressing such issues, as they can more accurately capture linguistic patterns, contextual nuances, and textual ambiguities than traditional models, representing a substantial advancement in improving label accuracy. This paper proposes utilizing LLMs to address humaninduced labeling bias in spam detection and applying different prompt design methods to guide the process. In text classification, we surveyed two leading-edge LLMs, ChatGPT and Gemini, and evaluated them on the English SMS spam dataset source from UC Irvine's Machine Learning Repository. We explored the highest-performing prompt designs using approaches like in-context learning. The findings indicate that in-context techniques for prompting improve model effectiveness by reducing human-induced (contextual) labeling bias in SMS spam detection with a Balanced Accuracy of 82% 97% and an Equal Opportunity Difference (EOD) of precisely zero, indicating LLMs' trustworthiness (fairness) in reducing this bias compared to traditional machine learning approaches. Our results also suggested that expanding the sample size can decrease LLMs' ability to reduce human-induced labeling bias in spam detection. In general, this study provides information on the strengths and limitations of LLMs and suggestions for methods to minimize human-induced labeling bias in spam detection

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

The Past, Present, and Future of Storage Technologies (Part 1 of 2)

引用

ACM TRANSACTIONS ON STORAGE 2025年第1期21卷 1-3页

作者： Kuenning, Geoff Won, Youjip Zhao, Ming Computer Science Harvey Mudd College Claremont United States Electrical Engineering Korea Advanced Institute of Science and Technology Daejeon Korea (the Republic of) School of Computing and Information Sciences Arizona State University Tempe United States Computer Science Stony Brook University Stony Brook United States

来源：评论

学校读者我要写书评

暂无评论

Mm-Wave Metamaterial Antenna Design with Superstrate for B5G FR3 Bands Using Osprey Optimization Algorithm

Mm-Wave Metamaterial Antenna Design with Superstrate for B5G...

引用

2025 IEEE International Workshop on Antenna Technology, iWAT 2025

作者： Rekkas, Vasileios P. Boursianis, Achilles D. Koulouridis, Stavros Zaharis, Zaharias D. Christodoulou, Christos G. Goudos, Sotirios K. School of Physics Aristotle University of Thessaloniki ELEDIA@AUTH Thessaloniki541 24 Greece University of Patras Department of Electrical and Computer Engineering Patra Greece School of Electrical and Computer Engineering Aristotle University of Thessaloniki ELEDIA@AUTH Thessaloniki541 24 Greece The University of New Mexico Department of Engineering and Computing New Mexico United States School of Physics Aristotle University of Thessaloniki ELEDIA@AUTH Thessaloniki Greece

ISBN: (纸本)9798331527365

In this work, we present an antenna design for operation between 21 and 23 GHz in the FR3 band for B5G and 6G operation. The proposed antenna uses a metamaterial super-strate for gain and bandwidth enhancement. We optimize the geometrical parameters using a recently proposed evolutionary algorithm, namely the Osprey Optimization Algorithm (OOA), which imitates the behavior of osprey in nature. The OOA finds a design with wideband operation and gain enchantment. The proposed low-profile antenna yields a peak gain of about 8.9 dBi around 22 GHz. The antenna exhibits a frequency band with a $S_{11} © 2025 IEEE.

关键词： Microwave antennas

来源：评论

学校读者我要写书评

暂无评论

Measuring biases in AI-generated co-authorship networks

引用

EPJ Data Science 2025年第1期14卷 1-33页

作者： Kalhor, Ghazal Ali, Shiza Mashhadi, Afra School of Electrical and Computer Engineering College of Engineering University of Tehran Tehran Iran Computing and Software Systems University of Washington Bothell USA

Large Language Models (LLMs) have significantly advanced prompt-based information retrieval, yet their potential to reproduce or amplify social biases remains insufficiently understood. In this study, we investigate this issue through the concrete task of reconstructing real-world co-authorship networks of computer science (CS) researchers using two widely used LLMs—GPT-3.5 Turbo and Mixtral 8x7B. This task offers a structured and quantifiable way to evaluate whether LLM-generated scholarly relationships reflect demographic disparities, as co-authorship is a key proxy for collaboration and recognition in academia. We compare the LLM-generated networks to baseline networks derived from DBLP and Google Scholar, employing both statistical and network science approaches to assess biases related to gender and ethnicity. Our findings show that both LLMs tend to produce more accurate co-authorship links for individuals with Asian or White names, particularly among researchers with lower visibility or limited academic impact. While we find no significant gender disparities in accuracy, the models systematically favor generating co-authorship links that overrepresent Asian and White individuals. Additionally, the structural properties of the LLM-generated networks differ from those of the baseline networks. These results highlight the importance of examining how LLMs represent social and scientific relationships, particularly in contexts where they are increasingly used for knowledge discovery and scholarly search.

关键词：

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：