检索结果-内蒙古大学图书馆

IEEE Students' Conference on electrical, Electronics and computer science (SCEECS)

作者： Hemal Singloo Manoj Kumar Simran Namdev Kinsukh Suni Sonbawne Nishant Indoria Dept. Computer Science and Engineering VIT Bhopal University Sehore India

This research paper investigates the effectiveness of deep learning models for gait recognition using a variety of data, including gait phase data and sensor data. This study evaluates the performance of convolutional neural networks (CNN) and long-term memory networks (LSTM), focusing on comparative analysis across multiple datasets to gain insight and contribute to different data analyses. The gait level dataset contains important movement patterns as a main area of research. Additionally, sensor-based data identifies various information needs, and complete analysis, for greater understanding and use. These comprehensive experimental results provide several observations on the variability and performance of CNNs and LSTMs across various datasets. The applicability of these models in recognizing activities in different products is presented, and their advantages and limitations in model recognition, display, and feature extraction are seen. This study distinguishes between different types of deep learning models and demonstrates their suitability for handling different types of data. These results lead to decisions regarding model selection and application, contributing to advances in analytics, biometrics, and human recognition applications.

关键词：

来源：评论

学校读者我要写书评

暂无评论

The Internet of Musical Things Meets Satellites: Evaluating Starlink Support for Networked Music Performances in Rural Areas

The Internet of Musical Things Meets Satellites: Evaluating ...

引用

Internet of Sounds, International Symposium on the

作者： Luca Turchet Paolo Casari Dept. of Information Engineering and Computer Science University of Trento Trento Italy

ISBN: (数字)9798350366525

ISBN: (纸本)9798350366532

Low Earth Orbit (LEO) satellite constellations offer a viable alternative to traditional fixed and wireless technologies, ensuring comparable speeds and latency even in rural areas. One of the applications that can be served by such types of satellite-based Internet access is represented by networked music performance (NMP) systems, which connect remote musicians and enable them to play together. Such systems are particularly relevant to communities located in rural areas, as they might lack access to local music education services, or may encounter difficulties in traveling to meet in presence for playing together. The integration of NMP system with LEO satellites could offer a viable alternative to in-person collaborative music activities, al-lowing musical stakeholders to reduce the costs for travel and the times to commute, with the benefit of zeroing pollution. However, thus far the use of NMP systems over satellite connections has been largely overlooked by the research community. To bridge this gap, we evaluate the performance of two Starlink-based communication infrastructures in supporting an NMP between two Italian cities, Trento and Verona. The first infrastructure connected two rural areas in the respective cities via two satellite links, one per terminal. The second connected one rural area in Trento to a neighborhood of Verona, respectively via a satellite link and a wired terrestrial network. Our findings clearly indicate that the performance of the current Starlink service in the tested geographical areas does not adhere to the strict latency and reliability requirements of NMPs. This calls for further research in this scarcely investigated domain at the frontiers of 6G networks.

关键词： Wireless communication Satellite constellations Satellites Pollution Urban areas Music Low earth orbit satellites Internet Stakeholders Reliability

来源：评论

学校读者我要写书评

暂无评论

Meta Reinforcement Learning for Resource Allocation in Multi-Antenna UAV Network with Rate Splitting Multiple Access

arXiv

引用

arXiv 2024年

作者： Zarini, Hosein Dehkordi, Maryam Farajzadeh Farhadi, Armin Mili, Mohammad Robat Movaghar, Ali Rasti, Mehdi Li, Yonghui Wong, Kai-Kit Dept. of Computer Engineering Sharif University of Technology Tehran Iran Dept. of Electrical and Computer Engineering George Mason University VA United States Dept. of Electrical and Computer Engineering University of Tehran Tehran Iran Tehran Iran Centre for Wireless Communications University of Oulu Oulu Finland School of Electrical and Information Engineering University of Sydney Sydney Australia Dept. of Electronic and Electrical Engineering University College London London United Kingdom

Unmanned aerial vehicles (UAVs) with multiple antennas have recently been explored to improve capacity in wireless networks. However, the strict energy constraint of UAVs, given their simultaneous flying and communication tasks, renders the exploration of energy-efficient multi-antenna techniques indispensable for UAVs. Meanwhile, lens antenna subarray (LAS) emerges as a promising energy-efficient solution that has not been previously harnessed for this purpose. In this paper, we propose a LAS-aided multi-antenna UAV to serve ground users in the downlink transmission of the terahertz (THz) band, utilizing rate splitting multiple access (RSMA) for effective beam division multiplexing. We formulate an optimization problem of maximizing the total system spectral efficiency (SE). This involves optimizing the UAV’s transmit beamforming and the common rate of RSMA. By recasting the optimization problem into a Markov decision process (MDP), we propose a deep deterministic policy gradient (DDPG)-based resource allocation mechanism tailored to capture problem dynamics and optimize its variables. Moreover, given the UAV’s frequent mobility and consequential system reconfigurations, we fortify the trained DDPG model with a meta-learning strategy, enhancing its adaptability to system variations. Numerically, more than 20% energy efficiency gain is achieved by our proposed LAS-aided multiantenna UAV equipped with 4 lenses, compared to a single-lens UAV. Simulations also demonstrate that at a signal-to-noise (SNR) of 10 dB, the incorporation of RSMA results in a 22% SE enhancement over conventional orthogonal beam division multiple access. Furthermore, the overall system SE improves by 27%, when meta-learning is employed for fine-tuning the conventional DDPG method in literature. © 2024, CC BY.

关键词： Markov processes

来源：评论

学校读者我要写书评

暂无评论

A Scratchpad Spiking Neural Network Accelerator

A Scratchpad Spiking Neural Network Accelerator

引用

Computing and Machine Intelligence (ICMI), International Conference on

作者： Rasha Karakchi Dept. Computer Science and Engineering University of South Carolina Columbia SC

ISBN: (数字)9798350372977

ISBN: (纸本)9798350372984

Spiking Neural Networks (SNNs) have recently been used as a computational model for applications such as deep learning, image recognition and machine learning. Similar to the biological brain, SNN neurons depend on the membrane level to fire an output. If the level exceeds a specified threshold, the neuron sends an output to activate next neurons. This leads to an unbalanced workload among the neurons. The dynamically-changing membrane level is stored inside a neuron. In hardware, this storage can be implemented as a register or on-chip memory, which determines the amount of consumed resources and, in turns, affects the network scalability. SNN accelerators have recently been implemented on UltraScale FPGAs devices for high-performance purposes. On-chip memories on these devices are classified as distributed memory, Block RAMs (BRAMs) and Ultra RAMs (URAMs). In this paper, we explored the impact of using different on-chip memories to store the membrane level of SNN neurons. We implemented a parameterizable SpIking Neural networK (SINK) accelerator where the network capacity and weight_width are parameters. SINK has the ability to run in four different modes based on the memory type. We ran SINK on UltraScale Zyncl04 FPGA device and measure the utilization of the hardware resources (LUTs), registers, memory, power consumption and performance. The results show that URAM can be the best fit to store the membrane level, since it used 30%, 11 % and 2% less LUTs, Regs, and power 2% respectively comparing with BRAM and distributed memory

关键词： Neurons Memory management Random access memory Spiking neural networks Biomembranes Hardware System-on-chip

来源：评论

学校读者我要写书评

暂无评论

On Fusing ChatGPT and Ensemble Learning in Discontinuous Named Entity Recognition in Health Corpora

arXiv

引用

arXiv 2024年

作者： Chen, Tzu-Chieh Lin, Wen-Yang Dept. of Computer Science and Information Engineering National University of Kaohsiung Taiwan

Named Entity Recognition (NER) has traditionally been a key task in natural language processing (NLP), aiming to identify and extract important terms from unstructured text data. However, a notable challenge for contemporary deep-learning NER models has been identifying discontinuous entities, which are often fragmented within the text. To date, methods to address Discontinuous Named Entity Recognition (DNER) have not been explored using ensemble learning to the best of our knowledge. Furthermore, the rise of large language models (LLMs, such as ChatGPT) in recent years has shown significant effectiveness across many NLP tasks. Most existing approaches, however, have primarily utilized ChatGPT as a problem-solving tool rather than exploring its potential as an integrative element within ensemble learning algorithms. In this study, we investigated the integration of ChatGPT as an arbitrator within an ensemble method, aiming to enhance performance on DNER tasks. Our method combines five state-of-the-art (SOTA) NER models with ChatGPT using custom prompt engineering to assess the robustness and generalization capabilities of the ensemble algorithm. We conducted experiments on three benchmark medical datasets, comparing our method against the five SOTA models, individual applications of GPT-3.5 and GPT-4, and a voting ensemble method. The results indicate that our proposed fusion of ChatGPT with the ensemble learning algorithm outperforms the SOTA results in the CADEC, ShARe13, and ShARe14 datasets, achieving improvements in F1-score of approximately 1.13%, 0.54%, and 0.67%, respectively. Compared to the voting ensemble method, our approach achieved improvements of about 0.63%, 0.32%, and 0.09%. Furthermore, compared to GPT-3.5 and GPT-4, our average results were approximately 7.42%, 0.89%, and 0.54% higher. The results demonstrate the effectiveness of our proposed fusion method of ChatGPT and ensemble algorithms, showcasing its potential to enhance NLP applications in

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Enhancing Pneumonia Detection: CNN Interpretability with LIME and SHAP

Enhancing Pneumonia Detection: CNN Interpretability with LIM...

引用

International Conference on electrical engineering and Information & Communication Technology (ICEEICT)

作者： Md. Hamid Hosen Arnob Saha Altaf Uddin Kahakashan Ashraf Sadia Nawar Dept. of Computer Science & Engineering East Delta University Chittagong Bangladesh

ISBN: (数字)9798350385779

ISBN: (纸本)9798350385786

A bacterial or viral infection of the lungs can cause pneumonia, one of the dangerous and potentially fatal illnesses that can have dire repercussions in a short amount of time. Therefore, a key component of a successful treatment plan is an early diagnosis. Therefore, a sophisticated and automated system that can diagnose chest X-rays and make the process of diagnosing pneumonia easier for both specialists and novices is required. This research aims to create a CNN model that will help with the accurate classification of pneumonia. In this work, we have presented our Deep Learning method for the classification challenge, which is taught using modified images through several pre-processing stages. With an overall accuracy of 93.30%, we were able to classify X-ray images of pneumonia using a custom CNN model. Our proposed model is able to precisely detect pneumonia from X-ray images with amazing accuracy and loss. Furthermore, we employed the LIME and SHAP tools of the XAI technique to generate a noteworthy conclusion to persuade medical practitioners.

关键词： Training Deep learning Pneumonia Computational modeling Predictive models Loss measurement Robustness

来源：评论

学校读者我要写书评

暂无评论

DevOps 2.0: Embracing AI/ML, Cloud-Native Development, and a Culture of Continuous Transformation

DevOps 2.0: Embracing AI/ML, Cloud-Native Development, and a...

引用

Pervasive Computing and Social Networking (ICPCSN), International Conference on

作者： V. Hemanth Das Chowdary A. Shanmukh T. Phaneendra Nikhil B. Sandeep Kumar Feroz Khan Dept. of Computer Science Engineering Koneru Lakshmiah University Vaddeswaram India

ISBN: (数字)9798350386349

ISBN: (纸本)9798350386356

This research study investigates the emerging domain of DevOps, assessing its capacity to fundamentally transform the processes of software development and delivery. A notable advancement in the field of DevOps is the amalgamation of AI/ML and artificial intelligence. These technologies possess significant potential to revolutionize the DevOps domain through the automation of tedious duties, including provisioning infrastructure, managing configurations, and identifying anomalies. This study highlights the increasing need for streamlined solutions. By harnessing the capabilities of AI and ML-driven tools, DevOps teams will have the ability to streamline their processes, enabling them to allocate resources to more strategic approaches such as nurturing innovation, addressing complex challenges, and making decisions based on data. This study will investigate the precise methods by which AI/ML can enhance DevOps practices, resulting in a software development environment that is more secure, efficient, and agile.

关键词： Industries Technological innovation DevOps Automation Social networking (online) Microservice architectures Transforms

来源：评论

学校读者我要写书评

暂无评论

Beta-Sigma VAE: Separating beta and decoder variance in Gaussian variational autoencoder

arXiv

引用

arXiv 2024年

作者： Kim, Seunghwan Lee, Seungkyu Dept. of Computer Science and Engineering Kyung Hee University Korea Republic of

Variational autoencoder (VAE) is an established generative model but is notorious for its blurriness. In this work, we investigate the blurry output problem of VAE and resolve it, exploiting the variance of Gaussian decoder and β of beta-VAE [14]. Specifically, we reveal that the indistinguishability of decoder variance and β hinders appropriate analysis of the model by random likelihood value, and limits performance improvement by omitting the gain from β. To address the problem, we propose Beta-Sigma VAE (BS-VAE) that explicitly separates β and decoder variance σx2 in the model. Our method demonstrates not only superior performance in natural image synthesis but also controllable parameters and predictable analysis compared to conventional VAE. In our experimental evaluation, we employ the analysis of rate-distortion curve and proxy metrics on computer vision datasets. The code is available on https://***/overnap/BS-VAE. Copyright © 2024, The Authors. All rights reserved.

关键词： Decoding

来源：评论

学校读者我要写书评

暂无评论

A Real-Time Detection of Patient's Aggressive Behaviors on a Smart Sensor

A Real-Time Detection of Patient's Aggressive Behaviors on a...

引用

Midwest Symposium on Circuits and Systems (MWSCAS)

作者： Yudai Inoue Vasily G. Moshnyaga Dept. Electronics Engineering and Computer Science Fukuoka University Fukuoka Japan

ISBN: (数字)9798350387179

ISBN: (纸本)9798350387186

Patient aggression in older people with dementia is one of the most challenging problems of medical care in today's nursing homes. Existing wireless autography-based technologies help caregivers to collect data related to behavioral and psychological symptoms of dementia. However, their inability to process the data in real-time reduces the efficiency of patient assessment and affects the quality of care. In this paper we present a computationally inexpensive technique that identifies aggressive behaviors on a sensor that measures the wrist acceleration of the dominant patient hand. Experiments show that the technique effectively detects aggressive behaviors in real-time and significantly reduces the wireless data transmissions.

关键词： Wireless communication Wrist Wireless sensor networks Accuracy Medical services Real-time systems Frequency measurement

来源：评论

学校读者我要写书评

暂无评论

Unveiling Predictive Factors in Apple Quality: Leveraging LIME, SHAP, and the Synergy of Machine Learning Models and Artificial Neural Networks

Unveiling Predictive Factors in Apple Quality: Leveraging LI...

引用

International Conference on electrical engineering and Information & Communication Technology (ICEEICT)

作者： Rituparna Chowdhury Rudra Das Fowzia Binta Faruk Ananna Arnob Saha Sadia Nawar Md. Hamid Hosen Dept. of Computer Science & Engineering East Delta University Chittagong Bangladesh

ISBN: (数字)9798350385779

ISBN: (纸本)9798350385786

In the realm of fruit quality assessment, the demand for accurate identification methods has grown significantly. Traditional measures, such as physical and chemical analyses, prove impractical due to their time-consuming nature and potential sample damage. To address this imperative need for precise and timely apple quality detection, this research explores the application of machine learning algorithms. Recognizing the drawbacks of traditional methods reliant on physical and chemical measures for predicting fruit maturity, non-destructive techniques, such as hyperspectral imaging and machine learning algorithms, are meticulously investigated. The study goes on to evaluate and compare the performance of various machine learning approaches, including Random Forest Classifier (RF), K-Nearest Neighbors Classifier (KNN), Support Vector Classifier (SVC), Extreme Gradient Boosting Classifier (XGB), Light Gradient Boosting Machine Classifier (LGBM), Artificial Neural Network (ANN), and Voting Classifier. Among these, the Artificial Neural Network emerges as the top performer, achieving an impressive accuracy rate of 92.87%. Following the thorough evaluation of the models, Explainable Artificial Intelligence (XAI) techniques, specifically LIME and SHAP, are employed. These techniques playa crucial role in interpreting and visualizing the intricate decision- making process of the models. The insights gained underscore the models' robustness while providing valuable information on the decision-making processes, contributing significantly to advancements in apple quality detection methodologies.

关键词： Adaptation models Machine learning algorithms Biological system modeling Static VAr compensators Artificial neural networks Boosting Vectors

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：