检索结果-内蒙古大学图书馆

41st International Conference on Machine Learning, ICML 2024

作者： Liu, Haitong Li, Qiang Wai, Hoi-To Department of Computer Science ETH Zurich Switzerland Department of System Engineering and Engineering Management The Chinese University of Hong Kong Hong Kong

This paper studies the performative prediction problem where a learner aims to minimize the expected loss with a decision-dependent data distribution. Such setting is motivated when outcomes can be affected by the prediction model, e.g., strategic classification. We consider a state-dependent setting where the data distribution evolves according to a controlled Markov chain. We focus on stochastic derivative free optimization (DFO) where the learner is given access to a loss function evaluation oracle with the above Markovian data. We propose a two-timescale DFO(λ) algorithm that features (i) a sample accumulation mechanism that utilizes every observed sample to estimate the gradient of performative risk, (ii) a two-timescale diminishing step size that balances the rates of DFO updates and bias reduction. Under a non-convex optimization setting, we show that DFO(λ) requires O(1/Ε3) samples (up to a log factor) to attain a near-stationary solution with expected squared gradient norm less than Ε. Numerical experiments verify our analysis. Copyright 2024 by the author(s)

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Use of Euclidean Distance to Improve Fingerprint Matching Accuracy 4th

Use of Euclidean Distance to Improve Fingerprint Matching Ac...

引用

4th International Conference on Innovations in Computational Intelligence and computer Vision, ICICV 2024

作者： Bhargava, Neeraj Yadav, Rajendra Choudhary, Usha Department of Computer Science School of Engineering System Sciences MDS University Rajasthan Ajmer India Department of Computer and Communication Engineering Manipal University Jaipur Rajasthan Jaipur India

ISBN: (纸本)9789819769919

Fingerprint authentication is the most sophisticated method of all biometric techniques and has been thoroughly verified through various applications. Fingerprint matching has been done using several fingerprint recognition techniques. The methods for matching fingerprints have changed significantly over time, but Euclidean distance is still widely used since it is efficient and straightforward. Euclidean distance is still used in minutiae-based fingerprint matching to produce accurate results. The computational geometry, image processing, and patter recognition fields all make substantial use of Euclidean distance, a fundamental distance matching approach. The purpose of this work is to highlight the importance of normalization in improving contrast, particularly in noisy fingerprint photographs where accuracy is critical. We analyzed how to make fingerprint matching more robots against a reference template by utilizing Euclidean distance between small spots. We also analyzed algorithm’s effectiveness in producing dependable and accurate fingerprint matching results using MATLAB. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.

关键词： Computational geometry

来源：评论

学校读者我要写书评

暂无评论

Leveraging Transfer Learning for Spatio-Temporal Human Activity Recognition from Video Sequences

引用

computers, Materials & Continua 2023年第3期74卷 5017-5033页

作者： Umair Muneer Butt Hadiqa Aman Ullah Sukumar Letchmunan Iqra Tariq Fadratul Hafinaz Hassan Tieng Wei Koh School of Computer Sciences Universiti Sains MalaysiaPenang1180Malaysia Department of Computer Science The University of ChenabGujrat50700Pakistan Department of Software Engineering and Information System Universiti Putra MalaysiaSelangor43400Malaysia

Human Activity Recognition(HAR)is an active research area due to its applications in pervasive computing,human-computer interaction,artificial intelligence,health care,and social ***,dynamic environments and anthropometric differences between individuals make it harder to recognize *** study focused on human activity in video sequences acquired with an RGB camera because of its vast range of real-world *** uses two-stream ConvNet to extract spatial and temporal information and proposes a fine-tuned deep neural ***,the transfer learning paradigm is adopted to extract varied and fixed frames while reusing object identification *** state-of-the-art pre-trained models are exploited to find the best model for spatial feature *** temporal sequence,this study uses dense optical flow following the two-stream ConvNet and Bidirectional Long Short TermMemory(BiLSTM)to capture *** state-of-the-art datasets,UCF101 and HMDB51,are used for evaluation *** addition,seven state-of-the-art optimizers are used to fine-tune the proposed network ***,this study utilizes an ensemble mechanism to aggregate spatial-temporal features using a four-stream Convolutional Neural Network(CNN),where two streams use RGB *** contrast,the other uses optical flow ***,the proposed ensemble approach using max hard voting outperforms state-ofthe-art methods with 96.30%and 90.07%accuracies on the UCF101 and HMDB51 datasets.

关键词： Human activity recognition deep learning transfer learning neural network ensemble learning spatio-temporal

来源：评论

学校读者我要写书评

暂无评论

One Day International Cricket Match Score Prediction Using Machine Learning Approaches 1

One Day International Cricket Match Score Prediction Using M...

引用

1st IEEE International Conference on Smart Power Control and Renewable Energy, ICSPCRE 2024

作者： Hossain, Md. Sabbir Rumpa, Ummay Khadiza Hossain, Md. Shakil Shovon, Sazzadul Islam Huque, Md. Tahmidul Hasan, Mahamudul East West University Department of Computer Science and Engineering Dhaka Bangladesh Central Michigan University Department of Information System MI United States Department of Computer Science and Engineering Dhaka Bangladesh

ISBN: (纸本)9798350377002

Cricket is the game that binds Bangladeshis to-gether. The datasets were split into different parts based on the time of the cricket matches to experiment with the results. We have used Random Forest Regression and Linear Regression to experiment with the results. We have worked on different sets of independent variables while working with different datasets but runs and wickets were constant features. The linear Regression technique gave 0.58 of r-squared metric and its MAE: 32.857046003557706, MSE: 1793.4360657965133, RMSE: 42.34897951304746 values indicate that the regression model is making predictions that are reasonably close to the actual ob-served values. However, there is still some room for improvement. The r-squared metric of 0.787 was derived from a relatively brief dataset, MAE: 20.48880467759807, MSE: 690.1167559064677, RMSE: 26.270073389818837 metrics indicate that the regression model is making predictions that are very close to the actual observed values. Random Forest Regression's r-squared metric resulted in 0.807 but the learning curve pointed out an overfit model. Thus, linear regression contained the best result. We have applied the regularisation technique to achieve a better result. The dataset was collected from Github and Kaggle, sizes are (350899, 15) and (4037, 9) respectively. © 2024 IEEE.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Large circuit models: opportunities and challenges

引用

Science China(Information Sciences) 2024年第10期67卷 25-66页

作者： Lei CHEN Yiqi CHEN Zhufei CHU Wenji FANG Tsung-Yi HO Ru HUANG Yu HUANG Sadaf KHAN Min LI Xingquan LI Yu LI Yun LIANG Jinwei LIU Yi LIU Yibo LIN Guojie LUO Hongyang PAN Zhengyuan SHI Guangyu SUN Dimitrios TSARAS Runsheng WANG Ziyi WANG Xinming WEI Zhiyao XIE Qiang XU Chenhao XUE Junchi YAN Jun YANG Bei YU Mingxuan YUAN Evangeline F.Y.YOUNG Xuan ZENG Haoyi ZHANG Zuodong ZHANG Yuxiang ZHAO Hui-Ling ZHEN Ziyang ZHENG Binwu ZHU Keren ZHU Sunan ZOU Huawei Noah’s Ark Lab School of Integrated Circuits Peking University Faculty of Electrical Engineering and Computer Science Ningbo University Department of Electronic and Computer Engineering Hong Kong University of Science and Technology Department of Computer Science and Engineering The Chinese University of Hong Kong School of Integrated Circuits Southeast University Huawei HiSilicon Peng Cheng Laboratory School of Computer Science Peking University School of Microelectronics State Key Laboratory of Integrated Chips and System Fudan University School of Artificial Intelligence Shanghai Jiao Tong University

Within the electronic design automation(EDA) domain, artificial intelligence(AI)-driven solutions have emerged as formidable tools, yet they typically augment rather than redefine existing methodologies. These solutions often repurpose deep learning models from other domains, such as vision, text, and graph analytics, applying them to circuit design without tailoring to the unique complexities of electronic circuits. Such an “AI4EDA” approach falls short of achieving a holistic design synthesis and understanding,overlooking the intricate interplay of electrical, logical, and physical facets of circuit data. This study argues for a paradigm shift from AI4EDA towards AI-rooted EDA from the ground up, integrating AI at the core of the design process. Pivotal to this vision is the development of a multimodal circuit representation learning technique, poised to provide a comprehensive understanding by harmonizing and extracting insights from varied data sources, such as functional specifications, register-transfer level(RTL) designs, circuit netlists,and physical layouts. We champion the creation of large circuit models(LCMs) that are inherently multimodal, crafted to decode and express the rich semantics and structures of circuit data, thus fostering more resilient, efficient, and inventive design methodologies. Embracing this AI-rooted philosophy, we foresee a trajectory that transcends the current innovation plateau in EDA, igniting a profound “shift-left” in electronic design methodology. The envisioned advancements herald not just an evolution of existing EDA tools but a revolution, giving rise to novel instruments of design-tools that promise to radically enhance design productivity and inaugurate a new epoch where the optimization of circuit performance, power, and area(PPA) is achieved not incrementally, but through leaps that redefine the benchmarks of electronic systems' capabilities.

关键词： AI-rooted EDA large circuit models (LCMs) multimodal circuit representation learning circuit optimization

来源：评论

学校读者我要写书评

暂无评论

Search Operations With Geolocation Estimation of Missing Persons Based on Real-time Drone Images

引用

Journal of Institute of Control, Robotics and systems 2024年第8期30卷 890-896页

作者： Lee, Joohyuk Lee, Ho Jun Arachchige, Sasanka Kuruppu Kim, Namyoung Heo, Hyeonjeong Lee, Kyuman Department of Robot and Smart System Engineering Kyungpook National University Korea Republic of School of Computer Science and Engineering Kyungpook National University Korea Republic of School of Space Engineering Sciences Kyungpook National University Korea Republic of

The use of drones in search-and-rescue missions allows us to easily search areas that are inaccessible to humans and enables rapid and efficient mission execution with minimal manpower. In this paper, we propose a search operation method that involves automatically recognizing missing persons based on real-time images captured by a camera mounted on a drone and estimating their geolocation information. Given a particular search area, we plan a flight path while taking into consideration a cost function with constraints. Using a deep-learning model trained using cropped, generated, and augmented data, we recognize missing persons through real-time images taken by the drone following the planned path. Additionally, we estimate the geolocation of the missing persons by coordinate-transforming the reference pixels of recognized objects in the image. Based on the estimated geolocation, we identify identical objects and count the total number of objects recognized during missions. We validate the proposed search method by completing a search-and-rescue challenge using a drone. © ICROS 2024.

关键词： Cost functions

来源：评论

学校读者我要写书评

暂无评论

Safe Return Home Solution Using a Smart Drone Scout

引用

Journal of Institute of Control, Robotics and systems 2024年第9期30卷 937-944页

作者： Heo, Hyeonjeong Lee, Hojun Lee, Kyuman Department of Robot and Smart System Engineering Kyungpook National University Korea Republic of School of Computer Science and Engineering Kyungpook National University Korea Republic of School of Space Engineering Sciences Kyungpook National University Korea Republic of

The women's safe return home service is short on operating hours and manpower. To address this social issue, this paper proposes an unmanned solution using a smart drone. When a user calls this system, it generates an initial safe path based on data from electronic tag wearers (i.e., obstacles) and map information. As the obstacles move, the safe path is re-planned to avoid them. The smart drone tracks and monitors the users returning home along the planned safe path at a certain distance and angle in real time. Since the raw GPS data from the user's app are unreliable, these are corrected to map-matching points for the drone to determine the points to track. Moreover, the smart drone identifies corners to avoid losing the monitored field of view at corners and adjusts its tracking accordingly. Flight test results validate the overall functionality and performance of the proposed solution, i.e., the proposed system improves user safety by effectively minimizing contact with electronic tag wearers while maintaining a certain distance and angle from the user by controlling the direction and speed of the drone. © ICROS 2024.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Android Based Arduino Learning Model by Implementing Multi Marker Augmented Reality

Android Based Arduino Learning Model by Implementing Multi M...

引用

2023 International Conference on Informatics engineering, Science and Technology, INCITEST 2023

作者： Adler, John Setiyadi, Dianto Br Ginting, Selvia Lorena Rinaldy Ginting, Yogie Wicaksono, Mochamad Fajar Universitas Komputer Indonesia Department Computer Engineering Bandung Indonesia Universitas Komputer Indonesia Department of Computer System Bandung Indonesia Universitas Komputer Indonesia Department Computer System Bandung Indonesia Universitas Riau Pekanbaru Faculty of Engineering Department of Mechanical Engineering Indonesia

ISBN: (纸本)9798350317862

This research aims to develop an interactive learning system based on Arduino Uno basic material. To realize this system, Augmented Reality technology is used which will explain each Arduino Uno pin function along with an explanation of each example of the basic circuit provided. This started with the Arduino Uno learning process which often caused damage to the Arduino or the components that were connected to the Arduino itself during the process of making the circuit. This problem can occur due to a lack of basic understanding of theory. The use of Augmented Reality technology is able to display objects in the form of Arduinos in virtual 3D, so that 3D objects can appear, markers are used to mark each object, the introduction of the Arduino Uno itself can be simulated by providing a real picture using the multi marker method where there is a main marker containing 3D objects in the Arduino series and supporting markers containing explanations of the series. Thus, learning is not limited to conventional learning and minimizes errors in trials. The method used to develop the system is the waterfall method. The waterfall method itself is part of implementing a design with a good and reasonable target. Based on the tests obtained, the Arduino Uno learning application is able to run well using multiple markers, 3D or 2D objects can be run on a smartphone, and black box testing and marker reading testing are carried out with different lighting levels and reading from different sides. © 2023 IEEE.

关键词： Smartphones

来源：评论

学校读者我要写书评

暂无评论

Drone-Assisted Load Distribution Framework for Traffic Optimization in IoT Networks 17

Drone-Assisted Load Distribution Framework for Traffic Optim...

引用

17th International Conference on COMmunication systems and NETworkS, COMSNETS 2025

作者： Soni, Priyanka Mense, Omkar Manohar Kanti Addya, Sourav National Institute of Technology Cloud and Smart System Services Lab Department of Computer Science and Engineering Karnataka India

ISBN: (纸本)9798331531195

Device connectivity has been redefined by the rapid development of the Internet of Things (IoT) technology, enabling diverse applications in areas such as smart cities, smart homes, and healthcare. These applications produce massive amounts of data, making it a significant challenge to offload tasks to cloud servers. However, the physical distance between devices often leads to latency issues for IoT systems, impacting time-sensitive applications. Fog computing tackles this issue by processing data closer to IoT devices. However, a high user density can strain the capacity of a macro base station, potentially leading to congestion. Drone Base Stations (DBSs) provide a flexible solution by acting as mobile relays to reduce latency and traffic burden. To address this issue, this work proposes a drone-assisted load distribution strategy in IoT networks. In our proposed approach, we deploy DBS efficiently using a greedy optimization technique. However, after the deployment of DBS, an optimal base station assignment framework dynamically connects user equipment to the most suitable base station, helping to minimize latency and enhance network performance. This dual-phase approach provides a scalable and practical solution for realtime IoT applications. Through extensive simulations, ensures that our proposed approach achieves a more balanced assignment compared to baseline algorithms. © 2025 IEEE.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Early Detection and Classification of Tomato Leaf Blight Disease Using Deep Neural Networks 2

Early Detection and Classification of Tomato Leaf Blight Dis...

引用

2nd IEEE International Conference on Data Science and Network Security, ICDSNS 2024

作者： Halder, Anangsha Guha Neogi, Subhrendu Dutta, Shibdas Khamaru, Koustov Bhattacharya, Arghya Roy, Shovan The Neotia University Department of Computer Science and Engineering Sarisha India Dcg Data-Core System India Department of Information Technology Kolkata India University of Engineering and Management Department of Computer Science and Engineering Kolkata India

ISBN: (纸本)9798350373110

The world population relies on agricultural plants for food, and illnesses reduce yield, but proper plant disease monitoring can help to resolve such issues. computer vision and machine learning methods can detect plant illnesses early, reducing their negative consequences and overcoming the drawbacks of continual human monitoring. After pre-processing of input images, the target region needs to be isolated. The proposed models analyzes images for detection and classification to identify edges, colors, and textures. This research has agricultural applications for early disease identification. This study examined ResNet, MobileNet, DenseNet, InceptionV3 and modified CNN model on the PlantVillage dataset to classify tomato illnesses. The F1 score, accuracy, sensitivity, and specificity can evaluate models. This paper evaluates these models using a diverse collection of tomato leaf pictures, contributing to automated disease detection systems and may be helpful for improving farm management and sustainability. © 2024 IEEE.

关键词： Fruits

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：