检索结果-内蒙古大学图书馆

Distribution-flexible subset quantization for post-quantizing super-resolution networks

Science China(Information Sciences) 2025年第3期68卷 163-180页

作者： Yunshan ZHONG Mingbao LIN Jingjing XIE Yuxin ZHANG Fei CHAO Rongrong JI Institute of Artificial Intelligence Xiamen University Key Laboratory of Multimedia Trusted Perception and Efficient Computing Ministry of Education of ChinaXiamen University Tencent Youtu Lab Department of Artificial Intelligence School of Informatics Xiamen University Peng Cheng Laboratory

This paper introduces distribution-flexible subset quantization(DFSQ), a post-training quantization method for super-resolution networks. Our motivation for developing DFSQ is based on the distinctive activation distributions of current super-resolution models, which exhibit significant variance across samples and channels. To address this issue, DFSQ conducts channel-wise normalization of the activations and applies distribution-flexible subset quantization(SQ), wherein the quantization points are selected from a universal set consisting of multi-word additive log-scale values. To expedite the selection of quantization points in SQ, we propose a fast quantization points selection strategy that uses K-means clustering to select the quantization points closest to the centroids. Compared to the common iterative exhaustive search algorithm, our strategy avoids the enumeration of all possible combinations in the universal set, reducing the time complexity from exponential to linear. Consequently, the constraint of time costs on the size of the universal set is greatly relaxed. Extensive evaluations of various super-resolution models show that DFSQ effectively improves performance even without fine-tuning. For example, for 4-bit EDSR×2 on the Urban benchmark, DFSQ obtains 0.242 dB PSNR gains.

关键词： super-resolution post-training quantization distribution-flexible subset quantization neural network

来源：评论

学校读者我要写书评

暂无评论

CryptoQFL: Quantum Federated Learning on Encrypted Data 4

CryptoQFL: Quantum Federated Learning on Encrypted Data

引用

4th IEEE International Conference on Quantum computing and Engineering, QCE 2023

作者： Chu, Cheng Jiang, Lei Chen, Fan Luddy School of Informatics Computing and Engineering Indiana University BloomingtonIN United States

ISBN: (纸本)9798350343236

Recent advancements in Quantum Neural Networks (QNNs) have demonstrated theoretical and experimental performance superior to their classical counterparts in a wide range of applications. However, existing centralized QNNs cannot solve many real-world problems because collecting large amounts of training data to a common public site is time-consuming and, more importantly, violates data privacy. Federated Learning (FL) is an emerging distributed machine learning framework that allows collaborative model training on decentralized data residing on multiple devices without breaching data privacy. Some initial attempts at Quantum Federated Learning (QFL) either only focus on improving the QFL performance or rely on a trusted quantum server that fails to preserve data privacy. In this work, we propose CryptoQFL, a QFL framework that allows distributed QNN training on encrypted data. CryptoQFL is (1) secure, because it allows each edge to train a QNN with local private data, and encrypt its updates using quantum homomorphic encryption before sending them to the central quantum server;(2) communication-efficient, as CryptoQFL quantize local gradient updates to ternary values, and only communicate non-zero values to the server for aggregation;and (3) computation-efficient, as CryptoQFL presents an efficient quantum aggregation circuit with significantly reduced latency compared to state-of-the-art approaches. © 2023 IEEE.

关键词： Data privacy

来源：评论

学校读者我要写书评

暂无评论

Driver-Drowsiness Detection System Using Deep Learning (CNN) 5

Driver-Drowsiness Detection System Using Deep Learning (CNN)

引用

5th IEEE International Conference on Artificial Intelligence in Engineering and Technology, IICAIET 2023

作者： Ridwan, Firman Hung, Lai Po Universiti Malaysia Sabah Faculty of Computing and Informatics Sabah Kota Kinabalu Malaysia

ISBN: (纸本)9798350304152

Drowsiness and fatigue of drivers are amongst the significant causes of road accidents. Every year, this factor increases the amounts of deaths and fatalities globally. When drivers are drowsy or fatigued, the frequency of yawning is different from those in the normal state. This behavior can determine whether the drivers are fatigued or not. It's critical to use technologies to create and build systems that can detect and monitor drivers' levels of attention throughout the driving process, whether they're alert or sleepy. In this study, A high-precision artificial intelligence model was developed using deep learning to detect driver drowsiness during driving to enhance road safety statistics. The model was trained using a dataset containing pictures of people yawning and blinking. The model is then integrated with mobile applications for ease of use with smartphones for drowsiness detection. The application includes authentication, camera stream, trip history, trip statistics, alarm and notification, and emergency modules. Similar systems proposed by prior researchers using hardware devices such as webcam, smartwatch, and many more to predict drowsiness. The results achieved model's accuracy ranges from 62-87%, depending on the frames input from the camera stream and device performance. The mAP and the average recall of the model were 0.55 and 0.662, respectively. The testing and evaluation phase is divided into three parts: unit testing, integration testing, and user acceptance testing. The results of unit testing and integration testing indicate that the application is working as intended, with high accuracy in detecting driver drowsiness. User acceptance testing is done using the System Usability Scale questionnaire, which shows that the application is perceived to be easy to use, effective, and satisfactory. The proposed system is expected to reduce the statistics of road accidents by detecting drowsy drivers during their driving trips, and it is widely and eas

关键词： Roads and streets

来源：评论

学校读者我要写书评

暂无评论

FazBoard: An AI-Educational Hybrid Teaching and Learning System 8th

FazBoard: An AI-Educational Hybrid Teaching and Learning Sys...

引用

8th Future Technologies Conference, FTC 2023

作者： BenMessaoud, Fawzi Bolchini, Davide Ash, Eliza Tseng, Ching-Min IU Luddy School of Informatics Computing and Engineering IUPUI Bloomington United States

ISBN: (纸本)9783031474538

FazBoard, an avant-garde educational platform, seamlessly integrates artificial intelligence with contemporary educational methodologies to foster a dynamic, adaptive, and collaborative learning ecosystem. The platform’s linchpin comprises two components: an agile digital canvas that simulates interactive teaching and learning spaces, and an AI Assistant, incarnated as a digital humanoid, available round-the-clock for responsive academic support. The digital canvas is designed to cultivate an immersive and versatile environment, emulating the interactions of traditional classrooms without the constraints of time or location. The AI Assistant excels in providing instantaneous responses to queries, aggregating valuable learning analytics, and streamlining administrative tasks - all of which contribute to curriculum refinement and enhanced pedagogical efficacy. Furthermore, FazBoard is aimed at bolstering student engagement by creating an inclusive learning milieu, through adaptive learning strategies that cater to the diverse educational needs of its users. This paper delves into the architecture, functionalities, and far-reaching applications of FazBoard, heralding it as a quintessential model of integrating AI into education, thereby shaping the trajectory of 21st-century educational practices. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Solving the Blood Assignment Problem for Hospital in Pontianak City, Indonesia using Ant Colony Optimization 10

Solving the Blood Assignment Problem for Hospital in Pontian...

引用

10th International Conference on Information Technology, Computer, and Electrical Engineering, ICITACEE 2023

作者： Pangestika, Menur Wahyu Aziz, Zalilah Abd Razali, Razulaimi Bin College of Computing Informatics and Mathematics Selangor Darul Ehsan Shah Alam Malaysia

ISBN: (纸本)9798350322729

Blood is considered a very valuable and rare commodity because there is no chemical process that can produce blood. In addition, blood also has a limited shelf life. When viewed from the demand side of blood, blood has a different number of requests for each day, week and month, while blood must always be available. The hospital blood bank is responsible for blood availability, safe transfusion, quality and adequate transfusion. The main challenge is how the blood bank can minimize the amount of blood wasted at the end of the time in the hospital blood bank due to overestimation. If there is blood that is wasted, it will be in vain, while the demand for blood occurs every day. In this study, Ant Colony Optimization (ACO) algorithm is applied to solve the problem of finding the best route, that is finding the route to get the highest number of rooms to be visited and what is the minimum amount of blood not used. In this study, samples were used for blood type O and blood type PRC. From the results of applying the ACO algorithm for Blood Assignment Problem (BAP), the minimum wastage value is 0 with the number of rooms visited being 8. By knowing the amount of blood wasted, this algorithm can be used to optimize blood supply so as to avoid wasting blood due to overestimation. © 2023 IEEE.

关键词： Hospitals

来源：评论

学校读者我要写书评

暂无评论

Numerical Solution of SIR Model with Time Delay for Transmission of Tuberculosis Disease 11

Numerical Solution of SIR Model with Time Delay for Transmis...

引用

11th IEEE Conference on Systems, Process and Control, ICSPC 2023

作者： Ali, Zaileha Md Abd Aziz, Farah Hanani Azmi, Nura Ain Batcha, Siti Rahimah Nordin, Noratika Sadikin, Zubaidah School of Mathematical Sciences College of Computing Informatics and Mathematics Shah Alam Malaysia

ISBN: (纸本)9798350340860

Despite Malaysia's efforts to combat tuberculosis (TB) through strategies like early detection, effective treatment, and public awareness campaigns, the disease remains a significant health challenge, especially in specific regions and among vulnerable populations. The Susceptible-Infected-Recovered (SIR) model is a valuable tool for comprehending TB transmission dynamics, particularly the interactions between susceptible, infected, and recovered individuals;however, accurately modeling TB transmission involves addressing the critical factor of time delay, which includes the recovery period and the duration for an infected person to transition to the recovered state, essential for capturing the true dynamics of TB spread. This study analyzes the effect of recovery time delay on the SIR model using Fifth - Order Runge-Kutta (RK5) method, solving a system of delay differential equations (DDEs). The results demonstrate that the SIR model with recovery time delay exhibits distinct patterns for the period of recovery. The findings highlight the influence of time delay on the dynamics of the SIR model for TB in Malaysia, particularly in the growth of the infected population and prolonging the recovery of individuals. © 2023 IEEE.

关键词： Dynamics

来源：评论

学校读者我要写书评

暂无评论

Benchmark DPC++ code and performance portability on heterogeneous architectures 46

Benchmark DPC++ code and performance portability on heteroge...

引用

46th ICT and Electronics Convention, MIPRO 2023

作者： Mijic, Nenad Davidovic, Davor Rüer Bošković Institute Centre for Informatics and Computing Zagreb Croatia

ISBN: (纸本)9789532331042

Source code portability is becoming increasingly important in the development of new solutions in HPC due to the wide diversification of hardware and heterogeneity of systems. With Intel's oneAPI suite of programming tools and the Data Parallel C++ compiler, a single source code containing both host and device code can leverage hardware architectures from different vendors. Using the compiler's interoperability, it can be linked to existing libraries such as MPI to run the program on a distributed memory system. In this paper we benchmark and analyze the performance that can be achieved with the Intel DPC++ compiler, using the distributed Cholesky QR2 algorithm as an example and comparing it with the native CUDA and C++ implementation. The analysis shows that the performance degradation when using SYCL is negligible when a smaller number of nodes are used, but with the cost that some additional self-made optimizations are required in SYCL code. © 2023 MIPRO Croatian Society.

关键词： Interoperability

来源：评论

学校读者我要写书评

暂无评论

Distributed intelligence for IoT-based smart cities: a survey

引用

Neural computing and Applications 2024年第27期36卷 16621-16656页

作者： Hashem, Ibrahim Abaker Siddiqa, Aisha Alaba, Fadele Ayotunde Bilal, Muhammad Alhashmi, Saadat Mehmood Department of Computer Science College of Computing and Informatics University of Sharjah Sharjah27272 United Arab Emirates School of Computing College of Arts and Sciences Universiti Utara Malaysia Kedah Sintok06010 Malaysia Department of Computer Science Federal College of Education Zaria Kaduna Tudun Wada810282 Nigeria Department of Pharmaceutical Outcomes and Policy University of Florida GainesvilleFL32610 United States Department of Computing and Information Systems School of Engineering and Technology Sunway University Selangor Petaling Jaya47500 Malaysia Department of Information Systems College of Computing and Informatics University of Sharjah Sharjah27272 United Arab Emirates

The remarkable miniaturization of Internet of Things (IoT)-based systems and the rise of distributed intelligence are promising research paradigms in the design of smart cities. IoT and distributed intelligence are conjoined. On the one hand, IoT provides a digital connection to everyday physical devices, and on the other hand, distributed intelligence allows these connected devices to work together in a distributed environment to solve complex problems through learning and reasoning. In this article, we investigate and report the recent premier advances in distributed intelligence for IoT-based smart cities. We use bibliometric techniques which provide essential insights into many studies to evaluate the literature including publication year, type, and citation status on distributed intelligence and its variations published between 2013 and 2023. We categorize and classify literature by devising a taxonomy based on essential parameters, technologies, approaches, applications, and implications with respect to distributed intelligence. Moreover, we provide recent ongoing advancements in distributed intelligence and IoT such as sustainability, blockchain, 5G technologies, AI, edge computing expansion and IoT expansion. We also highlight some of the use cases which include traffic management, energy management, waste management, public safety, and healthcare. We provide core layers and services for implementing distributed intelligence in IoT-based smart cities and present several research challenges and issues. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Smart city

来源：评论

学校读者我要写书评

暂无评论

An Investigation of the Effect of Weather on Hand-Foot Mouth Disease Incidence in Sarawak 4th

An Investigation of the Effect of Weather on Hand-Foot Mouth...

引用

4th International Conference on Advances in Computational Science and Engineering, ICACSE 2023

作者： Ibrahim, Ashraf Osman Chen, Wong Yat Robert, Joel Zeolick Yi, Chiew Cheng Majid, Mazlina Abdul Faculty of Computing and Informatics Universiti Malaysia Sabah Sabah Kota Kinabalu88450 Malaysia Creative Advanced Machine Intelligence Research Centre Faculty of Computing and Informatics Universiti Malaysia Sabah Jalan UMS Sabah Kota Kinabalu88400 Malaysia Faculty of Computing Universiti Malaysia Pahang Al-Sultan Abdullah Pekan Pahang Malaysia

ISBN: (纸本)9789819729760

Hand, foot, and mouth disease (HFMD) is consistently present in Malaysia, with a rising number of reported cases. Sabah has encountered multiple outbreaks of this disease. However, detailed information regarding the epidemiology and molecular traits of the viruses responsible for these outbreaks remains limited. This study used two data sets;one dataset was acquired from ***, containing HFMD across various Malaysian states. The second dataset was sourced from the National Center for Biotechnology Information (NCBI) and consists of HFMD statistics for Malaysia, including case counts and weather data. Both datasets were merged to create a comprehensive dataset that was specifically focused on examining the Malaysian state of Sarawak. This consolidated dataset contains a total of 20 at-tributes, forming the basis for the subsequent analyses conducted in this study. The study employed four distinct machine learning methods for regression analysis, comprising Random Forest, XGBoost, Decision Tree, and LASSO regression. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Plant diseases

来源：评论

学校读者我要写书评

暂无评论

Rielkies Nearby Sales Agents Finder Using Haversine Method

Rielkies Nearby Sales Agents Finder Using Haversine Method

引用

2023 IEEE International Conference on computing, ICOCO 2023

作者： Sulaiman, Mohd Suffian Aziz, Ahmad Ajad Azmi, Azri Derasit, Zuraidah Ismail, Azlan Nor Abdul Rahim, Siti Khatiiah College of Computing Informatics & Mathematics Universiti Teknologi Mara Malaysia Universiti Teknologi Malaysia Razak Faculty of Technology and Informatics Malaysia

ISBN: (纸本)9798350302684

With their new culinary products, Rielkies Choco jars, which come in a variety of flavors, the chocolate-based company Rielkies, which was created by one of Malaysia's most well-known actors, has been able to gain popularity in Malaysia. A system that enables Rielkies's customers to locate their preferred and reliable sales agents is crucial given the large number of people registered as their product sales agents in each state throughout Malaysia. Due to privacy and security concerns, many customers steer clear of doing business online. Many customers are also concerned with the sales representatives' delivery strategies because they often favor cash on delivery due to its security. The development of a web-based tool to locate salespeople nearby was suggested in this article. The tool would give customers or users access to the contact information of salespeople as well as information on their location and preferred method of delivery. The Haversine technique is chosen as the most appropriate algorithm to employ for this project to calculate the distance. The system's functions include locating local sales representatives for customers and managing sales representatives for admin to confirm the sales representatives to ensure they are reliable and secure. The result demonstrates that the system's functionality has been attained. To make their transactions more secure and simple than before, the project will assist customers in locating their favorite sales representatives. © 2023 IEEE.

关键词： Sales

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：