检索结果-内蒙古大学图书馆

A novel deep dual self-attention and Bi-LSTM fusion framework for Parkinson’s disease prediction using freezing of gait: a biometric application

引用

Multimedia Tools and Applications 2024年第33期83卷 80179-80200页

作者： Habib, Zeeshan Mughal, Muhammad Ali Khan, Muhammad Attique Hamza, Ameer Alturki, Nazik Jamel, Leila Department of Electrical Engineering HITEC University Taxila Pakistan Department of Computer Science HITEC University Taxila Pakistan Department of Computer Science and Mathematics Lebanese American University Beirut Lebanon Department of Information Systems College of Computer and Information Sciences Princess Nourah bint Abdulrahman University Riyadh Saudi Arabia

Parkinson’s disease (PD) disorder is caused by the imbalance of inhibitory dopamine and excitatory acetylcholine neurotransmitters, which causes hindrance in locomotion. Freezing of gait (FOG), tremors, and bradykinesia are the most debilitating and disconcerting symptoms of Parkinson’s Disease (PD). FOG episodes occur due to neurological control disorder and motorized impairments, which sternly hinder voluntary movement and forward locomotion. The computational intelligence techniques can effectively predict PD from the sensor data. To detect the FOG episodes often experienced by PD patients, this paper utilizes a 5G spectrum that operates at 4.8 GHz. The main idea of the proposed system is to extract the wireless channel characteristics information (containing amplitude variances) that can be integrated into a 5G communication system. This system collects data through wireless devices such as dipole antennas, network interface cards, and RF signal generators. Five human activities were performed: fast walking, slow walking, sitting on a chair, FOG episodes, and voluntary stopping. We proposed a novel deep dual self-attention and Bi-LSTM fusion architecture for PD prediction to train the collected dataset. Two architectures have been proposed, namely Deep Dual Attention Neural Network (D2ANN) and BiLSTM Neural Network (BNN), for the classification of PD using Chanel State Information (CSI) data. After that, the most prominent information is extracted using self-attention activation. A newly informed weighted fusion approach is proposed, and tree growth optimization is employed for the best feature selection. Finally, the selected features are classified using a multilayer full SoftMax neural network. The experimental process was conducted on FOG wireless sensor network data and obtained the highest accuracy of 98.66%. Compared to the results with state-of-the-art techniques, the proposed framework shows improved accuracy, recall, and precision rates. © The Author(s

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Evaluation-free Time-series Forecasting Model Selection via Meta-learning

引用

ACM Transactions on Knowledge Discovery from Data 2025年第3期19卷 1-41页

作者： Abdallah, Mustafa Rossi, Ryan A. Mahadik, Kanak Kim, Sungchul Zhao, Handong Bagchi, Saurabh Purdue University in Indianapolis IndianapolisIN United States Adobe Systems Inc San JoseCA United States Department of Electrical and Computer Engineering Purdue Univ West LafayetteIN United States

Time-series forecasting models are invariably used in a variety of domains for crucial decision-making. Traditionally these models are constructed by experts with considerable manual effort. Unfortunately, this approach has poor scalability while generating accurate forecasts for new datasets belonging to diverse applications. Without access to skilled domain-knowledge, one approach is to train all the models on the new time-series data and then select the best one. However, this approach is nonviable in practice. In this work, we develop techniques for fast automatic selection of the best forecasting model for a new unseen time-series dataset, without having to first train (or evaluate) all the models on the new time-series data to select the best one. In particular, we develop a forecasting meta-learning approach called AutoForecast that allows for the quick inference of the best time-series forecasting model for an unseen dataset. Our approach learns both forecasting models' performances over time horizon of the same dataset and task similarity across different datasets. The experiments demonstrate the effectiveness of the approach over state-of-the-art (SOTA) single and ensemble methods and several SOTA meta-learners (adapted to our problem) in terms of selecting better forecasting models (i.e., 2 × gain) for unseen tasks for univariate and multivariate testbeds. AutoForecast has also significant reduction in inference time compared to the naïve approach (doing inference using all possible models and then selecting the best one), with median of 42 × across the two testbeds. We release our meta-learning database corpus (348 datasets), performances of the 322 forecasting models on the database corpus, meta-features, and source codes for the community to access them for forecasting model selection and to build on them with new datasets and models which can help advance automating time-series forecasting problem. In our released database corpus, we unveil new traces

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

Deep Learning-Based Joint Channel Equalization and Symbol Detection for Air-Water Optoacoustic Communications

引用

IEEE Transactions on Cognitive Communications and Networking 2024年第3期11卷 1482-1492页

作者： Mahmud, Muntasir Younis, Mohamed Ahmed, Masud Choa, Fow-Sen University of Maryland Baltimore County Baltimore Department of Computer Science and Electrical Engineering BaltimoreMD United States University of Maryland Baltimore County Department of Information Systems BaltimoreMD United States

The optoacoustic effect is triggered by directing an optical signal in the air (using laser) to the surface of water, leading to the generation of a corresponding acoustic signal inside the water. Careful modulation of the laser signal would enable an innovative method for direct communication in air-water cross-medium scenarios experienced in many civil and military applications. In order to achieve a high data rate, a multilevel amplitude modulation scheme can be used to generate different acoustic signals to transmit multiple symbols. However, accurately demodulating these acoustic signals can be challenging due to multipath propagation within the harsh underwater environment, inducing inter-symbol interferences. This paper proposes a deep learning-based demodulation technique that uses a U-Net for signal equalization and a Residual Neural Network for symbol detection. In addition, fine-tuning at the receiver side is also considered to increase the demodulation robustness. The proposed deep learning model has been trained with our laboratory constructed dataset containing eight levels of optoacoustic signals captured from three different underwater positions. The model is validated using two datasets containing severe interference due to multipath-generated echoes and reverberations. The results show that our demodulation model achieves 96.6% and 91.7% accuracy for the two datasets, respectively, which significantly surpasses the 72.9% and 65.30% accuracy achieved by the conventional peak detection-based technique. © 2015 IEEE.

关键词： Intersymbol interference

来源：评论

学校读者我要写书评

暂无评论

Decision Tree-Based IoT Botnet Attack Detection 2

Decision Tree-Based IoT Botnet Attack Detection

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Al-Najjar, Hazem Al-Rousan, Nadia College Of Information Technology University Of Petra Department Of Computer Science Amman Jordan German Jordanian University Faculty Of Electrical Engineering And Information Systems Department Of Computer Engineering Jordan

ISBN: (纸本)9798350394962

This paper investigated the predictive capabilities of three decision tree models for IoT botnet attack prediction using packet information while minimizing the number of predictors. The study employed three decision tree models (C5, CHAID, and Random Forest) and two additional models (Logistic Regression and Bayesian Network) for comparison purposes. The IoT botnet attack dataset comprised various devices, mainly two types of attacks (Gafgyt and Mirai) and 23 feature engineering variables. Simulation results from all models showed that Accuracy, Precision, Recall, and False Omission Rate (FOR) values are approximately one, with an F1 score of around 0.5. CHAID and C5 models outperform other models in predicting IoT botnet attacks, as they are developed using only four and two variables, respectively. These results demonstrated that decision tree models with fewer variables can perform better than models that utilize all predictors. © 2024 IEEE.

关键词： Forecasting

来源：评论

学校读者我要写书评

暂无评论

Improving the Efficiency of Smart Grid Scheduling: A Linear Regression Model with Meta Job Information 2

Improving the Efficiency of Smart Grid Scheduling: A Linear ...

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Al-Najjar, Hazem Al-Rousan, Nadia College Of Information Technology University Of Petra Department Of Computer Science Amman Jordan German Jordianian University Faculty Of Electrical Engineering And Information Systems Department Of Computer Engineering Jordan

ISBN: (纸本)9798350394962

In order to forecast the run time of the jobs that were submitted, this research provides two linear regression prediction models that include continuous and categorical factors. A continuous predictor is built using the number of CPUs, average CPU speed, and memory size, whereas a categorical predictor is built using the user ID, group ID, and executable ID. According to the findings, the prediction rates for categorical and continuous predictors are 61% and 1%, respectively. sixty times better than the earlier models, which used continuous variables as a foundational model to determine task complexity and weight. The effectiveness of the categorical predictor in enhancing a job scheduling problem is then evaluated by combining it with three suggested job scheduling strategies. The suggested algorithms incorporated metrics - predicted run time, waiting time, and resource requirement - to select the smallest jobs. According to the results, Algorithm 3 performs better than earlier models in both performance metrics. The variation in total execution time and average waiting time is 1.14 to 1.76 and 4.5, respectively compared to previous models, Additionally, Algorithms 1 and 2 demonstrating superior performance in every scenario. © 2024 IEEE.

关键词： Grid computing

来源：评论

学校读者我要写书评

暂无评论

Enhancing Image Security with Rossler Attractor-Based Encryption 2

Enhancing Image Security with Rossler Attractor-Based Encryp...

引用

2nd International Conference on Cyber Resilience, ICCR 2024

作者： Al-Najjar, Hazem Al-Rousan, Nadia University Of Petra Faculty Of Information Technology Department Of Computer Science Amman Jordan German Jordanian University Faculty Of Electrical Engineering And Information Systems Department Of Computer Engineering Jordan

ISBN: (纸本)9798350394962

we introduced image encryption algorithms with high sensitivity, such that even a single alteration in a plain-text image would result in a complete transformation of the ciphered image. The first algorithm employed pixel values in a substitution approach, whereas the second algorithm utilized pixel values to generate chaotic function keys. Additionally, we employed a Rossler chaotic function with pixel data to generate chaotic values for the shuffling and substitution processes. After subjecting these algorithms to rigorous testing using standard metrics, our experiments revealed the efficiency of the algorithms in image encryption. Notably, even a minute change in the original image resulted in a vastly different ciphered image. These proposed algorithms underscored the effectiveness of integrating chaotic functions with pixel information, significantly enhancing image security. © 2024 IEEE.

关键词： Cryptography

来源：评论

学校读者我要写书评

暂无评论

Small Object Detection Without Attention for Aerial Surveillance 13th

Small Object Detection Without Attention for Aerial Surveil...

引用

13th International Conference on Information Technology and its Applications, CITA 2024

作者： Choi, Jehwan Nguyen, Duy-Linh Vo, Xuan-Thuy Jo, Kang-Hyun Intelligent Systems Laboratory Department of Electrical Electronic and Computer Engineering University of Ulsan Ulsan44610 Korea Republic of

ISBN: (纸本)9783031741265

This paper introduces the development of an essential deep-learning model for surveillance systems utilizing high-mounted CCTV or drones. Objects seen from elevated angles often look smaller and may appear at different angles compared to ground-level observations. To improve the detection of small objects, we propose a network incorporating an element-wise multiplication module based on the vanilla Vision Transformer (ViT) architecture. However, traditional transformer models need significant computational resources, which may not be practical for edge devices like CCTV cameras or drones. Therefore, we apply the Attention-Free Transformer (AFT) to reduce computational requirements enabling real-time operation on low-capacity devices. We validate the performance by combining ViT and AFT with the YOLOv5 real-time object detection model. Practical applicability is confirmed by implementing it on the low-capacity device named ODROID H3+. Validation datasets include Autonomous Driving Drone, VisDrone, AerialMaritime, and PKLot, all containing numerous small-sized objects. Experimental results on the VisDrone dataset show that YOLOv5 nano + AFT reduces parameter count by 4.6% while increasing accuracy by 1%, making it an efficient network. The model size is suitable for edge device implementation at 3.7 MB. Similarly, Aerial Maritime and PKLot datasets indicate a decreased amount of parameters and increased accuracy. Hence, the proposed deep learning model is applicable for aerial surveillance systems. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.

关键词： Drones

来源：评论

学校读者我要写书评

暂无评论

Development of a Bias Compensating Q-Learning Controller for a Multi-Zone HVAC Facility

引用

IEEE/CAA Journal of Automatica Sinica 2023年第8期10卷 1704-1715页

作者： Syed Ali Asad Rizvi Amanda J.Pertzborn Zongli Lin IEEE the Department of Electrical and Computer Engineering Tennessee Technological UniversityCookevilleTN 38505 USA the Mechanical Systems and Controls Group in the Building Energy and Environment Division at the National Institute of Standards and Technology(NIST) GaithersburgMD 20899 USA the Charles L.Brown Department of Electrical and Computer Engineering University of VirginiaCharlottesvilleVA 22904-4743 USA

We present the development of a bias compensating reinforcement learning(RL)algorithm that optimizes thermal comfort(by minimizing tracking error)and control utilization(by penalizing setpoint deviations)in a multi-zone heating,ventilation,and air-conditioning(HVAC)lab facility subject to unmeasurable disturbances and unknown *** is shown that the presence of unmeasurable disturbance results in an inconsistent learning equation in traditional RL controllers leading to parameter estimation bias(even with integral action support),and in the extreme case,the divergence of the learning *** demonstrate this issue by applying the popular Q-learning algorithm to linear quadratic regulation(LQR)of a multi-zone HVAC environment and showing that,even with integral support,the algorithm exhibits bias issue during the learning phase when the HVAC disturbance is unmeasurable due to unknown heat gains,occupancy variations,light sources,and outside weather *** address this difficulty,we present a bias compensating learning equation that learns a lumped bias term as a result of disturbances(and possibly other sources)in conjunction with the optimal control *** results show that the proposed scheme not only recovers the bias-free optimal control parameters but it does so without explicitly learning the dynamic model or estimating the disturbances,demonstrating the effectiveness of the algorithm in addressing the above challenges.

关键词： HVAC control optimal tracking Q-learning reinforcement learning(RL)

来源：评论

学校读者我要写书评

暂无评论

A Deep Learning Approach for Brain Tumor Classification

A Deep Learning Approach for Brain Tumor Classification

引用

2024 International Conference on computer and Applications, ICCA 2024

作者： Abdelmoniem, Aya Ebrahim, Gamal A. Abdellatif, Mohammad M. The British University in Egypt Faculty of Engineering Electrical Engineering Department Cairo Egypt Ain Shams University Faculty of Engineering Computer and Systems Engineering Department Cairo Egypt

ISBN: (纸本)9798350367560

The medical field is incredibly stressful for both specialists and patients, especially when dealing with life threatening cases. Brain tumor identification is one of the most challenging tasks even for experts in the field. Fortunately, the evolution of machine learning techniques and their flexibility to be integrated into many different fields, including tumor identification and classification has greatly simplified this task. For the physicians to accurately diagnose a patient, they must first have a lot of knowledge and years of experience in addition to going through many cases. Machine learning simplifies this process when it comes to learning and using this knowledge to decide without human intervention. If the data fed to the model is accurate, the decisions made will be more accurate. In this paper, an automated machine learning model is developed to identify brain tumors. Different datasets are utilized in the developed model. The accuracy achieved by the model lies between 60% and 70%. Another dataset is then used, and the model shows an accuracy of 99%, precision of 83%, F1 score of 77% and recall of 75%. © 2024 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Design of an RF-CMOS Switched-Capacitor Power Amplifier for NB-IoT RF Transceivers

Design of an RF-CMOS Switched-Capacitor Power Amplifier for ...

引用

2023 International Young Engineers Forum in electrical and computer engineering, YEF-ECE 2023

作者： Santos, Ana Isabel Pedro Oliveira, Joao NOVA School of Science and Technology Department of Electrical and Computer Engineering Caparica Portugal Centre of Technology and Systems CTS-UNINOVA Department of Electrical and Computer Engineering Caparica Portugal

ISBN: (纸本)9798350300420

The paper presents the implementation of a Switched Capacitor Power Amplifier (SCPA) to be integrated into a Narrowband Internet of Things (NB-IoT) Transceiver. The SCPA is designed to operate at a frequency of 0.9GHz and aiming the maximum output power allowed by the standard of 23dBm. All the blocks within the SCPA were developed using RF components of a standard 130nm CMOS process, with a 1.2V/2.4V supply voltage. Results show that the SCPA is able to produce a maximum output power of 15.61dBm with a maximum Power-Added Efficiency (PAE) of 26.52% and a Total Harmonic Distortion (THD) of 0.68%. The measured HD2 and HD3 are-70.23dBc and-43.41dBc, respectively. Additional Process, Voltage and Temperature (PVT) corners and Monte Carlo simulations indicate the SCPA operates properly when subject to different conditions and mismatches. Besides the SCPA, a 16 QAM modulation stage is designed in VerilogA in order to evaluate the performance of the RF PA when a sequence of symbols is transmitted. © 2023 IEEE.

关键词： Power amplifiers

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：