检索结果-内蒙古大学图书馆

NeuroProbe: A Performance Profiling Feed Forward Neural Network Simulator for Hardware Interaction Analysis

SN computer science 2025年第5期6卷 1-22页

作者： Ayyappan, Binu Kumar, G. Santhosh Department of Computer Science Cochin University of Science and Technology Kochi India Department of Artificial Intelligence and Data Science Rajagiri School of Engineering & Technology Kochi India

NeuroProbe is a simple neural network simulator designed by authors specifically for educational purposes focusing on simulating inference phase on a computationally capable embedded hardware, aiming to provide a deeper understanding of hardware interactions and performance improvement. This simulator incorporates performance profiling capabilities, allowing students and educators to analyze and optimize the performance of a simulated trained neural network. By simulating the intricate hardware interactions, NeuroProbe enables users to gain insights into how various components impact the overall performance of neural networks. Using this simulator, learners can experiment with different configurations and architectures, identify potential bottlenecks, and explore optimization strategies to enhance neural network performance before deploying the model on an Edge Computing environment. It also enriches computer science education by enabling learners to gain deeper insights into hardware interactions and performance improvement. By identifying performance bottlenecks, learners can refine their understanding of neural networks and develop skills in optimizing their models for enhanced performance. In conclusion, NeuroProbe empowers users to bridge the gap between theoretical knowledge and practical implementation of neural network on an embedded device, fostering a more comprehensive understanding of neural networks and their hardware requirements. Currently, it is focused solely on single-threaded CPU-based computation enabling SSE, AVX. At present, NeuroProbe does not support SIMD extensions such as AVX-FMA, AVX-512, multi-threaded computation, and accelerators such as NPU, TPU, and GPU. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： Hardware–software interactions Neural network simulator NeuroProbe Performance bottlenecks Performance profiling Simulations

来源：评论

学校读者我要写书评

暂无评论

Characterization of Memory Access in Deep Learning and Its Implications in Memory Management

引用

computers, Materials & Continua 2023年第7期76卷 607-629页

作者： Jeongha Lee Hyokyung Bahn Institute of Artificial Intelligence and Software Ewha UniversitySeoul03760Korea Department of Computer Science and Engineering Ewha UniversitySeoul03760Korea

Due to the recent trend of software intelligence in the Fourth Industrial Revolution,deep learning has become a mainstream workload for modern computer *** the data size of deep learning increasingly grows,managing the limited memory capacity efficiently for deep learning workloads becomes *** this paper,we analyze memory accesses in deep learning workloads and find out some unique characteristics differentiated from traditional ***,when comparing instruction and data accesses,data access accounts for 96%–99%of total memory accesses in deep learning workloads,which is quite different from traditional ***,when comparing read and write accesses,write access dominates,accounting for 64%–80%of total memory ***,although write access makes up the majority of memory accesses,it shows a low access bias of 0.3 in the Zipf ***,in predicting re-access,recency is important in read access,but frequency provides more accurate information in write *** on these observations,we introduce a Non-Volatile Random Access Memory(NVRAM)-accelerated memory architecture for deep learning workloads,and present a new memory management policy for this *** considering the memory access characteristics of deep learning workloads,the proposed policy improves memory performance by 64.3%on average compared to the CLOCK policy.

关键词： Memory access deep learning machine learning memory access memory management CLOCK

来源：评论

学校读者我要写书评

暂无评论

Stacked autoencoder with weighted loss function for intrusion detection in IoT application

引用

Multimedia Tools and Applications 2024年 1-29页

作者： Gangula, Rekha Vutukuru, Murali Mohan Kumar, M. Ranjeeth Department of Computer Science and Engineering Vaagdevi Engineering College Telangana BollikuntaWarangal506005 India Department of Computer Science and Engineering Koneru Lakshmaiah Education Foundation Vaddeswaram AP Guntur522502 India School of Computer Science & amp Artificial Intelligence SR University Anantasagar Telangana Warangal506371 India

The fast increase of network traffic in recent times causes significant detection of intrusions in Internet of Things (IoT) environments. Currently, Deep Learning (DL) models play a crucial role in cyber security for malicious identification and intrusion detection in IoT networks. The existing methods have drawbacks like overfitting, data imbalance, and not completely capturing complex dependencies and relationships among input and features that are significant for intrusion detection. To overcome these limitations, Stacked Autoencoder (SAE) with weighted loss function is proposed for effective Intrusion Detection System (IDS). The SAE includes weighted loss function to minimize the overfitting issue and finally, the One Class-Support Vector Machine (OCSVM) is used in the classification layer to classify the intrusions. The database used for intrusion detection in IoT environment are Bot-IoT and ToN-IoT databases, which undergo pre-processing by using standard scalar and min–max normalization to remove duplicates and inconsistent data in the databases. Then, the optimal features are selected by using firefly optimizer which selects active features values for classification. In the resulting phase, the stacked autoencoder obtains 99.99%, and 99.70% of classification accuracy on the Bot-IoT and ToN-IoT databases, respectively, which are superior while compared to the traditional autoencoder model. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Support vector machines

来源：评论

学校读者我要写书评

暂无评论

Peer Collaboration in DBLP Using Graph Convolutional Network

引用

SN computer science 2025年第1期6卷 1-7页

作者： Kumar, Chintoo Dar, Showkat Ahmad Batchu, Charishma Singamaneni, Kundan Sai Panguluri, Venkata Lakshmi Alekhya Department of Artificial Intelligence and Data Science GITAM Karnataka Bengaluru 562163 India Department of Computer Science and Engineering GITAM Karnataka Bengaluru 562163 India

The increasing popularity of Graph-based neural network architectures plays a pivotal role in providing promising results in applications, viz., Friendship networks, Co-authorship networks, Product recommendations, etc. Many machine learning tasks can be addressed when dealing with graph representation, i.e., graph classification, node classification, link prediction, graph embedding, etc. The traditional rule-based methods have many drawbacks in comparison to graph representation learning, such as it requires manual feature engineering and struggles with unseen cases (generalizability), etc., in the case of rule-based methods. At the same time, graph representation learning overcomes those issues profoundly. The advent of graph-based deep neural network architectures inspired us to complete a graph classification problem to explore clusters of researchers who worked together in academic publications. We built the model by using a graph convolutional network to solve a multi-classification problem. We conducted a comparative study to check the efficacy of graph representation learning by GCNs on the DBLP dataset. A comparative study shows that the built model outperforms the baseline models. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2025.

关键词： DBLP Graph convolutional network Graph mining Social networks

来源：评论

学校读者我要写书评

暂无评论

sEMG-based Sarcopenia risk classification using empirical mode decomposition and machine learning algorithms

引用

Mathematical Biosciences and engineering 2024年第2期21卷 2901-2921页

作者： Kumar, Konki Sravan Lee, Daehyun Jamsrandoj, Ankhzaya Soylu, Necla Nisa Jung, Dawoon Kim, Jinwook Mun, Kyung Ryoul Center for Artificial Intelligence Korea Institute of Science and Technology Seoul Korea Republic of KHU-KIST Department of Converging Science and Technology Graduate School Kyung Hee University Seoul Korea Republic of Department of Human Computer Interface and Robotics Engineering University of Science and Technology Daejeon Korea Republic of Department of Computer Science Ozyegin University Istanbul Turkey

Early detection of the risk of sarcopenia at younger ages is crucial for implementing preventive strategies, fostering healthy muscle development, and minimizing the negative impact of sarcopenia on health and aging. In this study, we propose a novel sarcopenia risk detection technique that combines surface electromyography (sEMG) signals and empirical mode decomposition (EMD) with machine learning algorithms. First, we recorded and preprocessed sEMG data from both healthy and at-risk individuals during various physical activities, including normal walking, fast walking, performing a standard squat, and performing a wide squat. Next, electromyography (EMG) features were extracted from a normalized EMG and its intrinsic mode functions (IMFs) were obtained through EMD. Subsequently, a minimum redundancy maximum relevance (mRMR) feature selection method was employed to identify the most influential subset of features. Finally, the performances of state-of-the-art machine learning (ML) classifiers were evaluated using a leave-one-subject-out cross-validation technique, and the effectiveness of the classifiers for sarcopenia risk classification was assessed through various performance metrics. The proposed method shows a high accuracy, with accuracy rates of 0.88 for normal walking, 0.89 for fast walking, 0.81 for a standard squat, and 0.80 for a wide squat, providing reliable identification of sarcopenia risk during physical activities. Beyond early sarcopenia risk detection, this sEMG-EMD-ML system offers practical values for assessing muscle function, muscle health monitoring, and managing muscle quality for an improved daily life and well-being. © 2024 American Institute of Mathematical sciences. All rights reserved.

关键词： Empirical mode decomposition

来源：评论

学校读者我要写书评

暂无评论

Modelling a Multi-Perceptron Q Network Layer for Analyzing Patient Condition in Healthcare Sector

Modelling a Multi-Perceptron Q Network Layer for Analyzing P...

引用

2025 International Conference on Multi-Agent Systems for Collaborative intelligence, ICMSCI 2025

作者： Karthi, M. Manirajan, S.P. Prince, A. Santhosh, R. Department of Artificial Intelligence and Data Science India Karpagam Academy of Higher Education Department of Computer Science and Engineering Coimbatore India

ISBN: (纸本)9798331509828

Chronic diseases like chronic respiratory diseases, diabetes, heart disease (HD) and cancer are the important causes of mortality globally. The diagnosis of Heart-related diseases with a variety of symptoms or characteristics is incredibly complex. Internet of Things (loT) embedded solution has improved with the increase in smart wearable devices. However, people who undergo heart attacks unexpectedly have poor survival rates. As a result, an Internet of Things based Multi-Perceptron Q-Network (MQN)-based monitoring program is suggested for cardiac patients to assist in identifying heart disease and administering medication appropriately. The technology based on MQN is used to securely encrypt this sensor data before sending it to the cloud. The encrypted and decrypted data is eventually done and MQN classifier is used to complete the classification process. The two categories of data that make up the classified outcomes are i) regular and ii) irregular. It indicates the individual's cardiac status, and if the results are peculiar, the doctor is notified via text so that the patient can receive treatment. When compared to current algorithms, the MQN for HD diagnosis exhibits better enhancement in the estimated investigational results. Furthermore, when contrasted against the current MQN, the introduced Multi-layered Perceptron (MLP) system facilitates safe data transmission yields the maximum security level of 96%, and is accomplished in the shortest amount of time for both decryption and encryption. © 2025 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Enhancing QoE of WebRTC-based Video Conferencing using Deep Reinforcement Learning 24

Enhancing QoE of WebRTC-based Video Conferencing using Deep ...

引用

24th Asia-Pacific Network Operations and Management Symposium, APNOMS 2023

作者： Ko, Kyungchan Ryu, Sangwoo Hong, James Won-Ki Department of Computer Science and Engineering POSTECH Korea Republic of Graduate School of Artificial Intelligence POSTECH Korea Republic of

ISBN: (纸本)9788995004395

The proportion of video traffic within the total internet traffic is steadily increasing and then video traffic already accounts for over half of the internet traffic. The increase in video traffic is due to the growing number of users for video-related services such as video streaming, live streaming, and video telephony. With the increasing users on video-related services, the importance of the quality of experience (QoE) for these services will become even more crucial in the future. Numerous studies to enhance the experience quality of video streaming have been conducted using adaptive bitrate (ABR) algorithms and artificial intelligence (AI). However, this work focuses on a more complex problem: improving the experience quality in multi-party, bidirectional communication scenarios such as video conferences. We propose a system that applies deep reinforcement learning (DRL) to the media server of a webRTC-based video conferencing system to allocate a bitrate’s video stream that suits the network conditions for users. The proposed method was implemented and evaluated, demonstrating great improvements. When the network conditions changed dynamically, the proposed approach achieved approximately 56.2% higher video bitrate compared to existing methods, resulting in a 24.7% enhancement in user experience quality. Copyright 2023 KICS.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A System Design for 3D Visualization of Underwater Multi-Target Tracking Based on Unity 3D and GMPHD Filter

A System Design for 3D Visualization of Underwater Multi-Tar...

引用

2025 International Conference on Electronics, Information, and Communication, ICEIC 2025

作者： Kim, Seungho Lee, Scott Uk-Jin Hanyang University Bio-Artificial Intelligence Department of Computer Science and Engineering Ansan Korea Republic of Hanyang University Department of Computer Science and Engineering Ansan Korea Republic of

ISBN: (纸本)9798331510756

This paper proposes a design for a visualization system for underwater multi-target tracking based on the Gaussian Mixture Probability Hypothesis Density (GMPHD) filter using Unity 3D engine. Traditional analysis methods for underwater tracking algorithms rely primarily on two-dimensional graphs and numerical data, making it challenging to intuitively understand the complex behavior of tracking systems in three-dimensional underwater environments. Our proposed system design provides a framework for 3D visualization of target movements, sonar sensor detection areas, and filter estimation results, along with an interactive interface for parameter adjustment. The proposed architecture consists of four main components: simulation core, visualization module, environment simulation, and user interface, which would enable researchers to analyze algorithm performance under various scenarios effectively. Through real-time visualization and parameter adjustment capabilities, the system is designed to offer immediate feedback on how different parameters affect tracking performance. This architectural approach aims to provide researchers with a more intuitive understanding of the GMPHD filter's operation and facilitate the optimization of tracking algorithms for underwater environments. The proposed system design demonstrates the potential of 3D visualization tools in advancing underwater multi-target tracking research and algorithm development. © 2025 IEEE.

关键词： Target tracking

来源：评论

学校读者我要写书评

暂无评论

360° High-Resolution Depth Estimation via Uncertainty-Aware Structural Knowledge Transfer

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on artificial intelligence 2024年第11期5卷 5392-5402页

作者： Cao, Zidong Ai, Hao Vasilakos, Athanasios V. Wang, Lin Grimstad4604 Norway Artificial Intelligence Thrust Guangzhou511453 China HKUST Department of Computer Science and Engineering SAR Hong Kong999077 Hong Kong

To predict high-resolution (HR) omnidirectional depth maps, existing methods typically leverage HR omnidirectional image (ODI) as the input via fully supervised learning. However, in practice, taking HR ODI as input is undesired due to resource-constrained devices. In addition, depth maps are often with lower resolution than color images. Therefore, in this article, we explore for the first time to estimate the HR omnidirectional depth directly from a low-resolution (LR) ODI, when no HR depth ground truth (GT) map is available. Our key idea is to transfer the scene structural knowledge from the HR image modality and the corresponding LR depth maps to achieve the goal of HR depth estimation without any extra inference cost. Specifically, we introduce ODI super-resolution (SR) as an auxiliary task and train both tasks collaboratively in a weakly supervised manner to boost the performance of HR depth estimation. The ODI SR task extracts the scene structural knowledge via uncertainty estimation. Buttressed by this, a scene structural knowledge transfer (SSKT) module is proposed with two key components. First, we employ a cylindrical implicit interpolation function (CIIF) to learn cylindrical neural interpolation weights for feature up-sampling and share the parameters of CIIFs between the two tasks. Then, we propose a feature distillation (FD) loss that provides extra structural regularization to help the HR depth estimation task learn more scene structural knowledge. Extensive experiments demonstrate that our weakly supervised method outperforms baseline methods, and even achieves comparable performance with the fully supervised methods. © 2024 IEEE.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

Optimizing Crop Yields with Multilayer Perceptrons: A Data-Driven Approach 8

Optimizing Crop Yields with Multilayer Perceptrons: A Data-D...

引用

8th International Conference on Inventive Systems and Control, ICISC 202

作者： Nivetha, N. Usharani, S. Department of Computer Science and Engineering Villupuram India Department of Artificial Intelligence and Machine Learning Villupuram India

ISBN: (纸本)9798350386578

Precision agriculture has become a major change in crop farming. It utilises cutting-edge technologies to maximise field-level management. Precision agriculture has completely transformed crop production by leveraging the latest developments to maximize field-level management. Predicting crop yields with accuracy helps farmers reduce their environmental impact, increase productivity, and make well-informed decisions. Accurate and timely insights are frequently lacking in traditional agricultural yield prediction approaches. The study offers a deep learning method for precisely predicting agricultural yields. Accurate crop yield forecasts assist farmers in minimizing their negative environmental effects, boosting productivity, and making educated choices. However, there are many obstacles because conventional agricultural yield prediction methods frequently need more timely and precise insights. Despite their success, several challenges still exist. These include handling heterogeneous data, dealing with missing values, and the complexity of capturing non-linear relationships in the data. To determine whether decision trees or Multi-Layer Perceptrons (MLP) are ideal in crop yield prediction, these models are compared with each other. Multi-layer perceptrons (MLP) are prominent among these techniques. Even though the MLP model was more accurate, decision trees also are relevant to the prediction process. This means have the capability of understanding multi-layer intra-data intricacies through their structure whereas decision trees may overfit on noisy data or grow too deep hence leading to many splits also known as being bushy unless they are pruned to reduce this bushiness. The study suggests a novel method for predicting agricultural productivity using a Machine learning model Decision Tree and Multi-Layer Perceptrons (MLP). A web interface is also created to enable smooth communication with the prediction model. Through the usage of this interface, farmers and agr

关键词： Prediction models

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：