检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zhao, Dong Zang, Qi Wang, Shuang Sebe, Nicu Zhong, Zhun School of Artificial Intelligence Xidian University Shaanxi China Department of Information Engineering and Computer Science University of Trento Italy School of Computer Science and Information Engineering Hefei University of Technology China School of Computer Science University of Nottingham NottinghamNG8 1BB United Kingdom

Presently, pseudo-labeling stands as a prevailing approach in cross-domain semantic segmentation, enhancing model efficacy by training with pixels assigned with reliable pseudo-labels. However, we identify two key limitations within this paradigm: (1) under relatively severe domain shifts, most selected reliable pixels appear speckled and remain noisy. (2) when dealing with wild data, some pixels belonging to the open-set class may exhibit high confidence and also appear speckled. These two points make it difficult for the pixel-level selection mechanism to identify and correct these speckled close- and open-set noises. As a result, error accumulation is continuously introduced into subsequent self-training, leading to inefficiencies in pseudo-labeling. To address these limitations, we propose a novel method called Semantic Connectivity-driven Pseudo-labeling (SeCo). SeCo formulates pseudo-labels at the connectivity level, which makes it easier to locate and correct closed and open set noise. Specifically, SeCo comprises two key components: Pixel Semantic Aggregation (PSA) and Semantic Connectivity Correction (SCC). Initially, PSA categorizes semantics into "stuff" and "things" categories and aggregates speckled pseudo-labels into semantic connectivity through efficient interaction with the Segment Anything Model (SAM). This enables us not only to obtain accurate boundaries but also simplifies noise localization. Subsequently, SCC introduces a simple connectivity classification task, which enables us to locate and correct connectivity noise with the guidance of loss distribution. Extensive experiments demonstrate that SeCo can be flexibly applied to various cross-domain semantic segmentation tasks, i.e. domain generalization and domain adaptation, even including source-free, and black-box domain adaptation, significantly improving the performance of existing state-of-the-art methods. The code is available at https://***/DZhaoXd/SeCo. © 2024 Neural information

关键词：

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-based Source Code Classification and Visualization of Decision Rationales 28th

Deep Learning-based Source Code Classification and Visualiza...

引用

28th International Conference on Knowledge Based and Intelligent information and engineering Systems, KES 2024

作者： Oeda, Shinichi Mitsumori, Seiya Department of Information and Computer Engineering National Institute of Technology Kisarazu College 11-1 Kiyomidaihigashi 2-chome Kisarazu City Chiba Japan Advanced Course of Control and Information Engineering National Institute of Technology Kisarazu College 11-1 Kiyomidaihigashi 2-chome Kisarazu City Chiba Japan

In programming education, it is crucial to provide instruction that is tailored to students' proficiency levels. For this purpose, an objective evaluation of each student's coding ability is essential. Additionally, it is necessary to understand the characteristics of source code written by both beginners and advanced students. Previous research has successfully converted the structural information of source code into graphs and assessed coding skills using deep learning, achieving high accuracy. However, it remains unclear which specific structural elements significantly influence these assessments. This study addresses this gap by transforming source code into abstract syntax trees and developing a model that uses Graph Convolutional Networks to categorize code as either beginner or advanced users based on learned structural information. Furthermore, we apply Integrated Gradients to visualize the decision-making basis of our model and elucidate the structural characteristics distinguishing source code written by beginner and advanced users. © 2024 The Authors.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Textile-Based Washable Multimode Capacitive Sensors for Wearable Applications

IEEE Journal on Flexible Electronics

引用

IEEE Journal on Flexible Electronics 2024年第10期3卷 445-453页

作者： Mehmood, Muhammad Qasim Adnan, Malik Zulfiqar, Muhammad Hamza Aljaloud, Khaled A. Sarwar, Rimsha Hussain, Rifaqat Alqahtani, Ali H. Alomainy, Akram Punjab Lahore54000 Pakistan Department of Biomedical Engineering University of Engineering and Technology Narowal Campus Lahore54890 Pakistan College of Engineering King Saud University Muzahimiyah Branch Riyadh11451 Saudi Arabia Antenna and Electromagnetics Research Group School of Electronic Engineering and Computer Science Queen Mary University of London LondonE1 4NS United States

Nowadays, textile-based sensors are of great interest because of the significance of intelligent and smart textiles in wearable applications because of textiles’ washability, flexibility, and durability. We developed conductive yarn-based textile sensors for wearable multimode human–machine interfaces (HMIs), breathing, and walking pattern detection. The low-cost sewing process is used to develop interdigitated capacitive (IDC) sensor patterns on shirts, masks, and shoe soles using ultrafine highly conductive thread. Four sensor-based touchpads (SBTPs) were developed on the shirt and showed multiple modes of operation based on the pressure of the finger touch. The multimode capacitive sensors-based HMI is connected to the laptop wirelessly to perform three different functions from each sensor. The sensors exhibit a sensitivity of 34.675, 29.440, and 25.789 pF/N at low, medium, and high touch pressure. The developed mask detects the breathing pattern of humans, whether it’s slow, normal, or fast. Shoe Insole developed sensors to see the walking pattern, either slow, normal, or running. The response and recovery time of the sensor system is 11 and 10 ms, respectively. Sensors tested for 20 000 detection cycles and responded stability. Also, the sensors responded accurately after washing with water and detergent water. Reported textile sensors are washable, flexible, stretchable, comfortable, and reusable, showing the practicality of proposed sensors for personalized healthcare, smart textiles, and electronic textiles (e-textiles). © 2022 IEEE.

关键词： Smart textiles

来源：评论

学校读者我要写书评

暂无评论

A systematic literature review to implement large language model in higher education: issues and solutions

引用

Discover Education 2025年第1期4卷 1-25页

作者： Guizani, Sghaier Mazhar, Tehseen Shahzad, Tariq Ahmad, Wasim Bibi, Afsha Hamam, Habib College of Engineering Alfaisal University Riyadh 11533 Saudi Arabia School of Computer Science National College of Business Administration and Economics Lahore 54000 Pakistan Department of Computer Science and Information Technology School Education Department Government of Punjab Layyah 31200 Pakistan Department of Computer Engineering COMSATS University Islamabad Sahiwal Campus Sahiwal 57000 Pakistan Department of Computing School of Arts and Creative Technology University of Greater Manchester Manchester BL15AB United Kingdom Department of Computer and Software Technology University of Swat Swat 19200 Pakistan Faculty of Engineering Uni de Moncton Moncton E1A3E9 NB Canada School of Electrical Engineering University of Johannesburg Johannesburg 2006 South Africa International Institute of Technology and Management (IITG) Av. Grandes Ecoles Libreville BP 1989 Gabon Bridges for Academic Excellence - Spectrum Tunis Tunisia

Artificial intelligence-driven Chatbots, especially large language models (LLMs) like GPT-4, represent significant progress in digital education. These models excel in mimicking human-like text and transforming learning and teaching methods. This study examines the development, application, and impact of LLMs in education. It highlights their role in automating instructional tasks and promoting personalized learning experiences. Despite integration concerns and ethical debates, LLMs showcase the potential of AI to improve educational practices. Our research concludes that LLMs offer transformative opportunities for education. However, their incorporation requires careful ethical considerations, data privacy measures, and a balance between human educators and AI technologies. The findings suggest strategies for integrating LLMs into educational frameworks to enhance learning outcomes while preserving educational integrity. © The Author(s) 2025.

关键词： Artificial intelligence Education Large language models Learning systems Natural language processing systems Neural networks

来源：评论

学校读者我要写书评

暂无评论

Innovative use of immobilized zinc oxide-impregnated activated carbon (ZnO@CB) for effective treatment of leachate: modeling and predictive assessment

引用

Environmental Science and Pollution Research 2025年 1-17页

作者： Singh, Kulbir Lohchab, Rajesh Kumar Goel, Gaurav Waziri, Sadiq Abdullahi Aguedal, Hakim Allab, Yacine Elmeliani, Mohamed El Amine Elaissaoui Iddou, Abdelkader Liu, Bing Terashima, Mitsuharu Kaswan, Suresh Haryana Ambala133207 India Department of Environmental Science & Engineering Guru Jambheshwar University of Science and Technology Haryana Hisar125001 India Department of Energy and Environment Thapar Institute of Engineering and Technology Punjab Patiala147004 India Department of Civil Engineering Indian Institute of Technology Punjab Ropar140001 India Department of Civil and Water Resources Engineering University of Maiduguri PMB 1069 Maiduguri Nigeria Department of Preparatory Classes Higher School of Saharan Agriculture Adrar01000 Algeria Bab Ezzouar16111 Algeria Department of Materials Engineering Faculty of Chemistry University of Sciences and Technology of Oran - Mohamed Boudiaf 1505 Box El M’naouer Oran31000 Algeria Terashima Laboratory Faculty of Environmental Engineering The University of Kitakyushu 1-1 Hibikino Fukuoka Wakamatsuku Kitakyushu808-0135 Japan Resources and Environment Innovation Institute School of Municipal and Environmental Engineering Shandong Jianzhu University Shandong250101 China Department of Computer Science and Engineering Chandigarh University Punjab Chandigarh140413 India

This study examined the viability of column method utilizing the immobilized zinc oxide-loaded activated carbon obtained from corncob (ZnO@CB) to treat the landfill leachate. Instrumental techniques like BET, FTIR, SEM–EDX, and XRD were applied for the characterization of the adsorbents. The break through curve (BTC) was evaluated by altering the flow rate, bed height, and initial concentration of NH3-N and COD. At 35 cm bed height with an initial level of 3264 mg-COD/L, the optimal adsorption capacity was observed to be 35.44 mg-COD/g. Meanwhile, the optimal NH3-N adsorption capacity was 4.81 mg-NH3-N/g at a flow @ 1 mL/min, with an initial concentration of 460 mg-NH3-N/L, and a bed height of 35 cm. Both NH3-N and COD adsorption exhibited a correlation coefficient higher than 0.98 as calculated by linear plots of bed depth service time (BDST) equations, indicating that the column structure model was appropriate. The results reveal that the performance of the adsorption process could be well predicted by artificial neural network (ANN) at 4, 7, and 1 neuron for input, middle, and output layers, with a mean absolute error of 0.0096 and 0.0093 for COD and NH3-N reduction, respectively. In the RF model, higher values of R2 (0.9876 for COD and 0.9874 for NH3-N) indicate the model accuracy. The regenerated adsorbent achieved 54.2% and 54.1% removal of COD and NH3-N and adsorbent usage was feasible for up to three cycles. Results of BDST, ANN, and RF models revealed that packed column with immobilized ZnO@CB adsorbent is an efficient method for treating landfill leachate, highlighting the potential of ZnO@CB for industrial applications. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Agro-waste Artificial neural network Bohart-Adams model Breakthrough curve Fixed bed column

来源：评论

学校读者我要写书评

暂无评论

Text Classification Accuracy Enhancement Using Deep Neural Networks

Text Classification Accuracy Enhancement Using Deep Neural N...

引用

2023 Al-Sadiq International Conference on Communication and Information technology, AICCIT 2023

作者： Najm, Mobajjal Kadhim Aljumaily, Mustafa S. Al-Khaleefa, Ahmed Ayal, A.M. Mohammed, Dheyaa Abdulameer Challoob, Aqeel Luaibi Shahid Chamran University of Ahvaz Faculty of Engineering Computer Engineering Department Ahvaz Iran University of Tennessee Electrical Engineering and Computer Science Department Tennessee United States University of Misan Faculty of Education Department of Physics Maysan Iraq University of Misan Maysan Iraq Imam Ja'afar Al-Sadiq University College of Information Technology Department of Computer Technical Engineering Al-Muthanna Iraq Department of Computer Techniques Engineering1 Baghdad Iraq

ISBN: (纸本)9798350341881

Machine-learning systems are part of the field of artificial intelligence that automatically learn models from data to make better decisions. Natural language processing (NLP), using object and learning approaches, performs well in statistical tasks, such as text classification or sentiment mining. Currently, in the field of natural language processing (NLP), various models, methods, and architectures have been proposed and a wide range of NLP tasks have been developed. Most natural language processing (NLP) problems can be formulated as classification problems (decide on the class of this object depending on the object and its context). In this field, machine learning methods and especially using neural network models are the most used. In this research, we propose a deep learning architecture for text classification that can influence model performance. The basic steps of this proposed method include text pre-processing (including sentence segmentation, punctuation, rooting, speech tagging, and vector conversion), text feature extraction, and LSTM algorithm-based classification model construction. The dataset used in this study is 20NewsGroup and IMDB. The evaluation results are provided in the form of MSE, Precision, Recall, Accuracy, and F-Measure evaluation criteria. © 2023 IEEE.

关键词： Neural network models

来源：评论

学校读者我要写书评

暂无评论

Digitally-defined ultrathin transparent wireless sensor network for room-scale imperceptible ambient intelligence

引用

npj Flexible Electronics 2024年第1期8卷 869-878页

作者： Yunxia Jin Mengxia Yu Dat T.Nguyen Xin Yang Zhipeng Li Ze Xiong Chenhui Li Yuxin Liu Yong Lin Kong John S.Ho Institute for Health Innovation and Technology National University of SingaporeSingapore 117599Singapore Department of Biomedical Engineering National University of SingaporeSingapore 119276Singapore Department of Electrical and Computer Engineering National University of SingaporeSingapore 117583Singapore Integrative Sciences and Engineering Programme National University of Singapore Graduate SchoolSingapore 119077Singapore The N.1 Institute for Health National University of SingaporeSingapore 117456Singapore Institute of Materials Research and Engineering Agency for Science Technology and ResearchSingapore 138634Singapore Department of Mechanical Engineering University of UtahSalt Lake CityUT 84112USA

Wireless and battery-free radio-frequency(RF)sensors can be used to create physical spaces that ambiently sense and respond to human *** such sensors ultra-flexible and transparent is important to preserve the aesthetics of living environments,accommodate daily activities,and functionally integrate with ***,existing RF sensors are unable to simultaneously achieve high transparency,flexibility,and the electrical conductivity required for remote room-scale ***,we report 4.5μm RF tag sensors achieving transparency exceeding 90%that provide capabilities in room-scale ambient wireless *** develop a laser-assisted water-based adhesion-reversion process to digitally realize computer-aided RF design at *** individually tagging multiple objects and regions of the human body,we demonstrate multiplexed wireless tracking of humanenvironment interactions and physiological signals at a range of up to *** radio-frequency identification sensors open opportunities for non-intrusive wireless sensing of daily living spaces for applications in health monitoring and elderly care.

关键词： network ambient transparent

来源：评论

学校读者我要写书评

暂无评论

3D magnetic field vector measurement by magneto-optical imaging

引用

Journal of the Magnetics Society of Japan 2022年第2期46卷 37-41页

作者： Sakaguchi, H. Wada, R. Oya. S. Matsumura, T. Saito, H. Ishibashi, T. Department of Materials Science and Technology Nagaoka University of Technology 1603-1 Kamitomioka Niigata Nagaoka940-2188 Japan Department of Mathematical Science and Electrical-Electronic-Computer Engineering Akita University 1-1 Tegata Gakuen-machi Akita010-8502 Japan

Magneto-optical (MO) imaging using MO imaging plates is a magnetic imaging technique that enables real-time measurements and is expected to be used for non-destructive testing and for observing the magnetic domains of magnetic materials. In this study, we propose a quantitative measurement method for three dimensional (3D) magnetic field vector measurements. The x-and y-components of the magnetic fields within the measured plane are calculated from the measured z-component by using a signal transformation method based on magnetic field transfer functions. Furthermore, the magnetic field distributions at different heights are also obtained from a one-shot image. In this paper, 3D magnetic field vector measurements are demonstrated for ferrite magnets and an electrical steel sheet. © 2022, Magnetics Society of Japan. All rights reserved.

关键词： Magnetic fields

来源：评论

学校读者我要写书评

暂无评论

Prediction of Power Consumption and Indoor Ambient Conditions by Artificial Neural Network and Long Short-Term Memory 6th

Prediction of Power Consumption and Indoor Ambient Condition...

引用

6th International Conference on Intelligent Computing and Optimization, ICO 2023

作者： Intaraumnauy, Tosapon Pimavilai, Nattavit Sirisumrannukul, Somporn Department of Electrical and Computer Engineering Faculty of Engineering King Mongkut’s Univerity of Technology North Bangkok 1518 Pracharat 1 Rd. Wongsawang Bangsue Bangkok Thailand

ISBN: (纸本)9783031501500

A data analysis model is proposed to predict the indoor ambient conditions and the power consumption of an air conditioner from the information obtained from air conditioner adjustment and ambient conditions. The ambient and usage data were collected from sensors and internet of things (IoT) devices installed in an occupied home. An artificial neural network (ANN) has been developed to predict the power consumption of the air conditioner after adjusting the air conditioner’s target temperature, fan mode setting, and ambient conditions. In addition, a long short-term memory (LSTM) has been developed to predict the indoor temperature and the humidity of the air conditioner when it is controlled. With the information on the upcoming indoor ambient conditions that are expected to occur, the data analysis model can also indicate the maximum energy savings while satisfying the constraint of the resident’s comfort. The experimental results demonstrate that the proposed data analysis model can reliably predict the ambient conditions in the air-conditioned room and the power consumption of the air conditioner from the air-conditioning adjustment and ambient, giving the best-operating settings in any different environment scenarios and therefore shows potential for home energy savings for smart home applications. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

Prototype Development of Electric Vehicle Database Platform Using Serverless and Microservice Architecture with Intelligent Data Analytics 6th

Prototype Development of Electric Vehicle Database Platform ...

引用

6th International Conference on Intelligent Computing and Optimization, ICO 2023

作者： Sirisumrannukul, Somporn Piamvilai, Nattavit Limprapassorn, Sirawich Wattanakitkarn, Tanachot Loedchayakan, Touchakorn Department of Electrical and Computer Engineering Faculty of Engineering King Mongkut’s University of Technology North Bangkok Pracharat 1 Rd. Wongsawang Bangsue Bangkok1518 Thailand

ISBN: (纸本)9783031503290

To fully realize the benefits of EVs, this paper proposes a prototype platform that can perform essential functions as an EV data center for storing and centralizing the data acquired from EVs and chargers directly. Using serverless and microservice architecture, this platform is purposely designed and built for scalability and includes intelligent data analytics capabilities with two different tasks: calculation of CO2 emission, and intraday spatial charging demand analysis. The results confirm that the communication between EVs, chargers, and the platform using the REST API protocol satisfies the response time of the Open Charge Point Protocol (OCPP) 2.0.1 standard, and the backend performance is flexible in terms of scalability and database optimization to accommodate the number of EVs. Given real-time data, the data analytics functions can display CO2 emissions, and predict an intraday power demand. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Data Analytics

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：