The escalating challenge of traffic congestion in urban areas necessitates innovative solutions at the intersection of computer science and artificial intelligence. The main problem is in the outdated nature of conven...
详细信息
Continual learning and multi-task learning are commonly used machine learning techniques for learning from multiple tasks. However, existing literature assumes multi-task learning as a reasonable performance upper bou...
详细信息
Continual learning and multi-task learning are commonly used machine learning techniques for learning from multiple tasks. However, existing literature assumes multi-task learning as a reasonable performance upper bound for various continual learning algorithms, without rigorous justification. Additionally, in a multi-task setting, a small subset of tasks may behave as adversarial tasks, negatively impacting overall learning performance. On the other hand, continual learning approaches can avoid the negative impact of adversarial tasks and maintain performance on the remaining tasks, resulting in better performance than multi-task learning. This paper introduces a novel continual self-supervised learning approach, where each task involves learning an invariant representation for a specific class of data augmentations. We demonstrate that this approach results in naturally contradicting tasks and that, in this setting, continual learning often outperforms multi-task learning on benchmark datasets, including MNIST, CIFAR-10, and CIFAR-100.
During clinical decision-making, the segmentation process Is considered an important task because it provides valuable information about the location, size, and characterization of the brain tumor. However, manual hum...
详细信息
ISBN:
(数字)9798350351484
ISBN:
(纸本)9798350351491
During clinical decision-making, the segmentation process Is considered an important task because it provides valuable information about the location, size, and characterization of the brain tumor. However, manual human segmentation is error-prone, time-consuming, and requires skilled *** resonance imaging (MRI) could give extremely detailed images for the investigation and diagnosis of glioblastoma brain *** compared evaluated approaches on the BRATS 2021 and BRATS 2022 datasets and found that they outperformed and could compete with state-of-the-art algorithms in comparable *** our research, we focused on two crucial tasks: segmentation and MGMT classification. This study also addresses asn objective evaluation through performance evaluation of cutting-edge DL-based techniques for MR image analysis (Brats 2021-Brats 2022). Based on the findings of the contrasted methods, we can confirm that using a combination of DL techniques will produce more accurate segmentation results than depending on a single, unique methodology. For the second task, five distinct deep learning-based methods were evaluated to predict the methylation state of the MGMT promoter.
This paper presents an investigation into the application of Brain-Computer Interface (BCI) technology for emoji selection by analyzing electroencephalographic (EEG) signals. EEG data was captured to interpret neural ...
详细信息
ISBN:
(数字)9798350352931
ISBN:
(纸本)9798350352948
This paper presents an investigation into the application of Brain-Computer Interface (BCI) technology for emoji selection by analyzing electroencephalographic (EEG) signals. EEG data was captured to interpret neural activity, with a Support Vector Machine (SVM) classifier employed for accurate emoji recognition. The system demonstrated a classification accuracy of 93.28% while implementing advancedsignalprocessing techniques that reduced noise and improved real-time performance. This research highlights the viability of thought-driven emoji selection and its implications for advancing accessible communication technologies and immersive virtual environments.
In recent years, the demand for electricity of the whole society has been increasing rapidly, and the power suppliers have been unable to meet the power consumption demand of all users. In order to effectively balance...
详细信息
ISBN:
(数字)9798350376548
ISBN:
(纸本)9798350376555
In recent years, the demand for electricity of the whole society has been increasing rapidly, and the power suppliers have been unable to meet the power consumption demand of all users. In order to effectively balance the relationship between the power suppliers and users, and ensure the satisfaction of suppliers and the experience of users, a load optimization model is proposed. Based on Time-of-use tariff, the model comprehensively considers the income satisfaction of the power supplier, the satisfaction of the consumer's electricity expenditure and the satisfaction of electricity consumption, and takes Time-of-use tariff and elastic load as decision variables. Finally, the multi-objective function of residential user price response under Time-of-use tariff mechanism is constructed, and the genetic algorithm is used to optimize the model under various constraints. The experimental results show that the load optimization algorithm can effectively reduce the load curve of residential electricity, and can carry out peak cutting and valley filling on the premise of ensuring the satisfaction of users and power supplier.
Examinations are pivotal milestones in educational settings, serving not just as a measure of academic performance but also as a complex behavioral phenomenon. While traditional studies have explored student behaviors...
详细信息
ISBN:
(数字)9798350361186
ISBN:
(纸本)9798350361193
Examinations are pivotal milestones in educational settings, serving not just as a measure of academic performance but also as a complex behavioral phenomenon. While traditional studies have explored student behaviors during examinations, the intricate, high-dimensional nature of such behaviors often escapes conventional analytical approaches. This paper introduces a novel methodology that employs Distributed Machine Learning techniques to analyze student behavior during exams compre- hensively. We design a system architecture that captures various behavioral metrics, and apply distributed machine learning algorithms to dissect this multi-faceted data. The findings offer actionable insights into optimizing examination processes and provide a pioneering framework for leveraging advanced com- putational techniques in educational psychology. Our empirical results demonstrate a substantial improvement in understanding student behaviors, potentially influencing educational policies and pedagogical strategies.
Deep neural networks (DNNs) have set new standards in identifying and classifying irregular patterns in ECG (electrocardiogram) signals, surpassing previous methods. Despite the easy access and affordability of ECG se...
详细信息
ISBN:
(纸本)9783031646072;9783031646089
Deep neural networks (DNNs) have set new standards in identifying and classifying irregular patterns in ECG (electrocardiogram) signals, surpassing previous methods. Despite the easy access and affordability of ECG sensors, a critical bottleneck remains the limited availability of reliable data for complex heart rhythms like second and third-degree atrioventricular block, ventricular tachycardia, and supraventricular tachycardia. This shortage has been a significant obstacle to improving DNN algorithms. Recent studies have turned to Generative Adversarial Networks (GANs) to create synthetic ECG data, enhancing the diversity of training datasets. However, much of this research has only managed to produce basic ECG components, missing the intricate details found in real patient data that includes multiple heartbeats. Our research has taken a groundbreaking approach by converting ECG signals into a two-dimensional format, allowing us to utilize advanced GAN models originally developed for image processing. This method has enabled us to generate extended, realistic ECG sequences closely mimicking those from actual patients. We have tested and refined our model using two databases, Physionet and Chapman, and have successfully produced 10-second ECG sequences showcasing a variety of heart rhythms previously unachieved in other studies. Our innovative technique not only surpasses existing methods in generating high-quality, realistic ECG data but also sets a new benchmark in ECG synthesis.
This research centers on the development of an advanced model for analyzing and forecasting elderly health monitoring data, leveraging the Internet of Things (IoT) technology. A comprehensive health monitoring system ...
详细信息
ISBN:
(数字)9798331528676
ISBN:
(纸本)9798331528683
This research centers on the development of an advanced model for analyzing and forecasting elderly health monitoring data, leveraging the Internet of Things (IoT) technology. A comprehensive health monitoring system specifically designed for the elderly is established, enabling the real-time acquisition and transmission of daily health metrics through the integration of IoT technologies. During the data analysis phase, an association rule mining algorithm is utilized to identify hidden patterns and correlations within the elderly health data, thereby facilitating the prediction of health outcomes. A subsequent health prediction model is formulated, incorporating data mining and machine learning methodologies, with its effectiveness and reliability rigorously assessed through simulation experiments. The outcomes of these simulations demonstrate a significant enhancement in both the precision and efficiency of the prediction model. This progress provides a solid scientific basis for the health management of the elderly population. Further model simulations confirmed the accuracy and practicality of the developed prediction model, underscoring its extensive potential for application in elderly health management.
Modeling methods and algorithms for analog-to-digital conversion of information in control systems are significant components in the development and design of devices for analog signals digitizing. This process involv...
详细信息
ISBN:
(数字)9798350350043
ISBN:
(纸本)9798350350050
Modeling methods and algorithms for analog-to-digital conversion of information in control systems are significant components in the development and design of devices for analog signals digitizing. This process involves several steps, including sampling, quantization, and encoding, to convert continuous signals into a digital format that can be easily manipulated and analyzed by digital *** quest to enhance the speed of converters is driven by the ever-increasing demand for rapid and efficient data processing. The research into additive successive approximation methods has yielded a novel approach: bitwise balancing. This innovative method, supported by a robust mathematical model, promises to revolutionize analog-to-digital conversion by significantly reducing the conversion time. The empirical analysis suggests that this reduction could range from 6 to $25 \%$, marking a substantial improvement over existing methods. This breakthrough shows not only the potential for faster data processing but also sets a new direction for the development of technical apparatuses designed for swift information form conversion.
Ensuring secure transmission and storage of digital information is critical for any organization to function properly. To address this issue, encryption algorithms are commonly used. advanced Encryption Standard (AES)...
详细信息
Ensuring secure transmission and storage of digital information is critical for any organization to function properly. To address this issue, encryption algorithms are commonly used. advanced Encryption Standard (AES) has been globally adopted as the mainstay cipher algorithm for securing transmission networks and storage devices due to its easy implementation and compatibility with both hardware and software applications. Another reason for its widespread popularity is its uncompromisable nature against existing brute force attacks, making it practically unbreakable on existing computing power. AES implementation for battery operated devices requires an algorithm with low power consumption and high-speed encryption/decryption of digital data. This paper proposes an FPGA implementation of a high throughput parallel pipelined 128-bit AES algorithm with a low power key expansion mechanism for iterative stages. A 128-bit symmetric key has been used for undertaking 10 rounds of transformations. All the encryption and decryption transformations are simulated using iterative design methodology in order to minimize hardware consumption. Xilinx Artix-7 FPGA device is used for hardware evaluation and Verilog HDL for programming. Simulation and synthesis task has been performed on Xilinx Vivado v2021.1 IDE. The results exhibit high-rate encryption of 68 Gb/s and low energy consumption of 7 pJ/bit.
暂无评论