Search Results (5,407)

Recently, AI systems such as autonomous driving and smart homes have become integral to daily life. Intelligent multi-sensors, once limited to single data types, now process complex text and image data, demanding faster and more accurate processing. While integrating NPUs and sensors has improved processing speed and accuracy, challenges like low resource utilization and long memory latency remain. This study proposes a method to reduce processing time and improve resource utilization by virtualizing NPUs to simultaneously handle multiple deep-learning models, leveraging a hardware scheduler and data prefetching techniques. Experiments with 30,000 SA resources showed that the hardware scheduler reduced memory cycles by over 10% across all models, with reductions of 30% for NCF and 70% for DLRM. The hardware scheduler effectively minimized memory latency and idle NPU resources in resource-constrained environments with frequent context switching. This approach is particularly valuable for real-time applications like autonomous driving, enabling smooth transitions between tasks such as object detection and route planning. It also enhances multitasking in smart homes by reducing latency when managing diverse data streams. The proposed system is well suited for resource-constrained environments that demand efficient multitasking and low-latency processing. Full article

Detecting and tracking the preclinical stages of Alzheimer’s disease (AD) is now of particular interest due to the aging of the world’s population. AD is the most common cause of dementia, affecting the daily lives of those afflicted. Approaches in development can accelerate the evaluation of the preclinical stages of AD and facilitate early treatment and the prevention of symptom progression. Shifts in P300 amplitude and latency, together with neuropsychological assessments, could serve as biomarkers in the early screening of declines in cognitive abilities. In this study, we investigated the ability of the P300 indices evoked during a visual oddball task to differentiate pre-clinically diagnosed participants from normal healthy adults (HCs). Two preclinical stages, named asymptomatic AD (AAD) and prodromal AD (PAD), were included in this study, and a total of 79 subjects participated, including 35 HCs, 22 AAD patients, and 22 PAD patients. A mixed-design ANOVA test was performed to compare the P300 indices among groups during the processing of the target and non-target stimuli. Additionally, the correlation between these neurophysiological variables and the neuropsychological tests was evaluated. Our results revealed that neither the peak amplitude nor latency of P300 can distinguish AAD from HCs. Conversely, the peak latency of P300 can be used as a biomarker to differentiate PAD from AAD and HCs. The correlation results revealed a significant relationship between the peak latency of P300 and memory domain tasks, showing that less time-demanding neuropsychological assessments can be used. In summary, our findings showed that a combination of P300 latency and memory-requiring tasks can be used as an efficient biomarker to differentiate individuals with AAD from HCs. Full article

Solid-state batteries (SSBs) hold the potential to revolutionize energy storage systems by offering enhanced safety, higher energy density, and longer life cycles compared with conventional lithium-ion batteries. However, the widespread adoption of SSBs faces significant challenges, including low charge mobility, high internal resistance, mechanical degradation, and the use of unsustainable materials. These technical and manufacturing hurdles have hindered the large-scale commercialization of SSBs, which are crucial for applications such as electric vehicles, portable electronics, and renewable energy storage. This study systematically reviews the global SSB patent landscape using a cross-sectional bibliometric and thematic analysis to identify innovations addressing key technical challenges. The study classifies innovations into key problem and solution areas by meticulously examining 244 patents across multiple dimensions, including year, geographic distribution, inventor engagement, award latency, and technological focus. The analysis reveals significant advancements in electrolyte materials, electrode designs, and manufacturability. This research contributes a comprehensive analysis of the technological landscape, offering valuable insights into ongoing advancements and providing a roadmap for future research and development. This work will benefit researchers, industry professionals, and policymakers by highlighting the most promising areas for innovation, thereby accelerating the commercialization of SSBs, and supporting the transition toward more sustainable and efficient energy storage solutions. Full article

In this paper, we present a modular, high-performance prototype platform for real-time event extraction, designed to address key challenges in processing large volumes of unstructured data across applications like crisis management, social media monitoring and news aggregation. The prototype integrates advanced natural language processing (NLP) techniques (Term Frequency–Inverse Document Frequency (TF-IDF), Latent Semantic Indexing (LSI), Named Entity Recognition (NER)) with data mining strategies to improve precision in relevance scoring, clustering and entity extraction. The platform is designed to handle real-time constraints in an efficient manner, by combining TF-IDF, LSI and NER into a hybrid pipeline. Unlike the transformer-based architectures that often struggle with latency, our prototype is scalable and flexible enough to support various domains like disaster management and social media monitoring. The initial quantitative and qualitative evaluations demonstrate the platform’s efficiency, accuracy, scalability, and are validated by metrics like F1-score, response time, and user satisfaction. Its design has a balance between fast computation and precise semantic analysis, and this can make it effective for applications that necessitate rapid processing. This prototype offers a robust foundation for high-frequency data processing, adaptable and scalable for real-time scenarios. In our future work, we will further explore contextual understanding, scalability through microservices and cross-platform data fusion for expanded event coverage. Full article

Efficient BCIs (Brain-Computer Interfaces) harnessing EEG (Electroencephalography) have shown potential in controlling mobile robots, also presenting new possibilities for assistive technologies. This study explores the integration of advanced deep learning models—ASTGCN, EEGNetv4, and a combined CNN-LSTM architecture—with ROS (Robot Operating System) to control a two-wheeled mobile robot. The models were trained using a published EEG dataset, which includes signals from subjects performing thought-based tasks. Each model was evaluated based on its accuracy, F1-score, and latency. The CNN-LSTM architecture model exhibited the best performance on the cross-subject strategy with an accuracy of 88.5%, demonstrating significant potential for real-time applications. Integration with ROS was facilitated through a custom middleware, enabling seamless translation of neural commands into robot movements. The findings indicate that the CNN-LSTM model not only outperforms existing EEG-based systems in terms of accuracy but also underscores the practical feasibility of implementing such systems in real-world scenarios. Considering its efficacy, CNN-LSTM shows a great potential for assistive technology in the future. This research contributes to the development of a more intuitive and accessible robotic control system, potentially enhancing the quality of life for individuals with mobility impairments. Full article

With generative Artificial Intelligence (AI) capturing public attention, the appetite of the technology sector for larger and more complex Deep Learning (DL) models is continuously growing. Traditionally, the focus in DL model development has been on scaling the neural network’s foundational structure to increase computational complexity and enhance the representational expressiveness of the model. However, with recent advancements in edge computing and 5G networks, DL models are now aggressively being deployed and utilized across the cloud–edge–IoT continuum for the realization of in situ intelligent IoT services. This paradigm shift introduces a growing need for AI practitioners, as a focus on inference costs, including latency, computational overhead, and energy efficiency, is long overdue. This work presents a benchmarking framework designed to assess DL model scaling across three key performance axes during model inference: classification accuracy, computational overhead, and latency. The framework’s utility is demonstrated through an empirical study involving various model structures and variants, as well as publicly available datasets for three popular DL use cases covering natural language understanding, object detection, and regression analysis. Full article

The proliferation of 5G networks has revolutionized wireless communication by delivering enhanced speeds, ultra-low latency, and widespread connectivity. However, in heterogeneous cloud radio access networks (H-CRAN), efficiently managing inter-cell interference while ensuring energy conservation remains a critical challenge. This paper presents a novel energy-efficient, dynamic enhanced inter-cell interference coordination (eICIC) scheme based on deep reinforcement learning (DRL). Unlike conventional approaches that focus primarily on optimizing parameters such as almost blank subframe (ABS) ratios and bias offsets (BOs), our work introduces the transmission power during ABS subframes (TPA) and the channel quality indicator (CQI) threshold of victim user equipments (CTV) into the optimization process. Additionally, this approach uniquely integrates energy consumption into the scheme, addressing both performance and sustainability concerns. By modeling key factors such as signal-to-interference-plus-noise ratio (SINR) and service rates, we introduce the concept of energy-utility efficiency to balance energy savings with quality of service (QoS). Simulation results demonstrate that the proposed scheme achieves up to $70\%$ energy savings while enhancing QoS satisfaction, showcasing its potential to significantly improve the efficiency and sustainability of future 5G H-CRAN deployments. Full article

The integration of blockchain and federated learning (FL) has emerged as a promising solution to address data privacy and security challenges in Intelligent Unmanned Ports (IUPs). However, existing blockchain federated learning (BFL) frameworks encounter significant limitations, including high latency, inefficient data processing, and limited scalability, particularly in scenarios with sparse and distributed data. This paper introduces a novel multi-chain federated learning (MFL) framework to overcome these challenges. The proposed MFL architecture interconnects multiple BFL chains to facilitate the secure and efficient aggregation of data across distributed devices. The framework enhances privacy and efficiency by transmitting aggregated model updates rather than raw data. A low-frequency consensus mechanism is employed to improve performance, leveraging game theory for representative selection to optimize model aggregation while reducing inter-chain communication overhead. The experimental results demonstrate that the MFL framework significantly outperforms traditional BFL in terms of accuracy, latency, and system efficiency, particularly under the conditions of high data sparsity and network latency. These findings highlight the potential of MFL to provide a scalable and secure solution for decentralized learning in IUP environments, with broader applicability to other distributed systems such as the Industrial Internet of Things (IIoT). Full article

With the development of Internet of Vehicles (IoV) technology, the need for real-time data processing and communication in vehicles is increasing. Traditional request-based methods face challenges in terms of latency and bandwidth limitations. Mode 4 in cellular vehicle-to-everything (C-V2X), also known as autonomous resource selection, aims to address latency and overhead issues by dynamically selecting communication resources based on real-time conditions. However, semi-persistent scheduling (SPS), which relies on distributed sensing, may lead to a high number of collisions due to the lack of centralized coordination in resource allocation. On the other hand, non-orthogonal multiple access (NOMA) can alleviate the problem of reduced packet reception probability due to collisions. Age of Information (AoI) includes the time a message spends in both local waiting and transmission processes and thus is a comprehensive metric for reliability and latency performance. To address these issues, in C-V2X, the waiting process can be extended to the queuing process, influenced by packet generation rate and resource reservation interval (RRI), while the transmission process is mainly affected by transmission delay and success rate. In fact, a smaller selection window (SW) limits the number of available resources for vehicles, resulting in higher collisions when the number of vehicles is increasing rapidly. SW is generally equal to RRI, which not only affects the AoI part in the queuing process but also the AoI part in the transmission process. Therefore, this paper proposes an AoI estimation method based on multi-priority data type queues and considers the influence of NOMA on the AoI generated in both processes in C-V2X system under different RRI conditions. Our experiments show that using multiple priority queues can reduce the AoI of urgent messages in the queue, thereby providing better service about the urgent message in the whole vehicular network. Additionally, applying NOMA can further reduce the AoI of the messages received by the vehicle. Full article

Background/Objectives: To examine the relationships between state affect and sleep health at within- and between-person levels among health science students. Methods: A correlational design was used and 54 health science students were included. The participants completed baseline and 7-day ambulatory assessments in a free-living setting. Daily sleep and affect were measured using the Consensus Sleep Diary and Positive and Negative Affect Schedule. Mixed-effect models were used to examine the effects of affect on sleep health. Results: The participants were 19.8 (SD, 0.6) years and 92.6% were females. Approximately 40% had poor sleep quality. Controlling for the potential confounders (e.g., age, sex, and bedtime procrastination), higher within-person negative affect predicted shorter sleep duration, lower sleep efficiency, longer sleep onset latency, and less feeling rested. Higher between-person negative affect predicted shorter sleep duration. Higher within-person positive affect predicted longer sleep onset latency. Higher within- and between-person positive affect predicted more feeling rested. Conclusions: Negative affect was most consistently associated with sleep health at the individual level. Affect regulation should be considered when delivering personalized interventions targeting sleep health among health science students. Full article

Remote direct memory access (RDMA) is widely used within and across data centers due to its low latency, high throughput, and low CPU overhead. To further enhance the transmission performance of RDMA, techniques such as multi-path RDMA have been proposed. However, while these techniques increase throughput, they also introduce significant out-of-order (OoO) packet issues that standard RDMA network interface cards (RNICs) struggle to handle effectively. To address the OoO challenges in RDMA network and ensure the integrity of data, we propose an FPGA-based bitmap which is capable of maintaining high throughput and low latency under OoO conditions. Our design segments the bitmap and maintains status information, achieving the low-latency processing of OoO packets with excellent scalability, thus making it suitable for various network environments. We implement this design on Xilinx AU200 FPGA and test it in a simulated 100 Gbps data center network. The results show that the performance under OoO transmission conditions is comparable to that under in-order conditions, demonstrating the solution’s effectiveness in handling RDMA OoO packets efficiently and ensuring high-performance data transfer in RDMA networks. Full article

Detecting anomalies in distributed systems through log analysis remains challenging due to the complex temporal dependencies between log events, the diverse manifestation of system states, and the intricate causal relationships across distributed components. This paper introduces a TLAN (Temporal Logical Attention Network), a novel deep learning framework that integrates temporal sequence modeling with logical dependency analysis for robust anomaly detection in distributed system logs. Our approach makes three key contributions: (1) a temporal logical attention mechanism that explicitly models both time-series patterns and logical dependencies between log events across distributed components, (2) a multi-scale feature extraction module that captures system behaviors at different temporal granularities while preserving causal relationships, and (3) an adaptive threshold strategy that dynamically adjusts detection sensitivity based on system load and component interactions. Extensive experiments on a large-scale synthetic distributed system log dataset show that TLAN outperforms existing methods by achieving a 9.4% improvement in F1-score and reducing false alarms by 15.3% while maintaining low latency in real-time detection. The framework demonstrates particular effectiveness in identifying complex anomalies that involve multiple interacting components and cascading failures. Through comprehensive empirical analysis and case studies, we validate that TLAN can effectively capture both temporal patterns and logical correlations in log sequences, making it especially suitable for modern distributed architectures. Our approach also shows strong generalization capability across different system scales and deployment scenarios, supported by thorough ablation studies and performance evaluations. Full article

In mobile networks, a content server can provide multiple services simultaneously to a mobile device, generating multiple data flows. As the device moves, the transmission path in the wired network may need to be switched to maintain service continuity. However, a single switching path may not be able to accommodate all the flows, potentially leading to congestion and a degraded user experience. To address this challenge, we propose a multi-topology routing-based mobile data scheduling method that dynamically switches flows across multiple paths to enhance flexibility and load balancing. The performance of this method is significantly influenced by the construction of logical topologies. Well-designed topologies provide high-bandwidth, low-latency paths to all possible destination nodes, while poorly designed topologies waste switch capacity and fail to achieve these goals. In this paper, we introduce an efficient multi-topology construction method for scheduling mobile data flows in software-defined networking (SDN). Our approach optimizes and balances transmission capacity for each destination node while adhering to the flow entry constraints of switches. Simulations demonstrate that our method consistently outperforms the single-path switching method and the other two multi-topology construction methods in terms of packet delay, packet loss rate, and network throughput, regardless of the device’s new location. Full article

Sensor networks generate vast amounts of data in real-time, which challenges existing predictive maintenance frameworks due to high latency, energy consumption, and bandwidth requirements. This research addresses these limitations by proposing an edge-cloud hybrid framework, leveraging edge devices for immediate anomaly detection and cloud servers for in-depth failure prediction. A K-Nearest Neighbors (KNNs) model is deployed on edge devices to detect anomalies in real-time, reducing the need for continuous data transfer to the cloud. Meanwhile, a Long Short-Term Memory (LSTM) model in the cloud analyzes time-series data for predictive failure analysis, enhancing maintenance scheduling and operational efficiency. The framework’s dynamic workload management algorithm optimizes task distribution between edge and cloud resources, balancing latency, bandwidth usage, and energy consumption. Experimental results show that the hybrid approach achieves a 35% reduction in latency, a 28% decrease in energy consumption, and a 60% reduction in bandwidth usage compared to cloud-only solutions. This framework offers a scalable, efficient solution for real-time predictive maintenance, making it highly applicable to resource-constrained, data-intensive environments. Full article

Background/Objectives: This study evaluated the antinociceptive effect of the Ximenia americana L. bark extract (HEXA) and its primary component, caffeic acid (CA), through in vivo assays. Methods: The antinociceptive properties were assessed using abdominal writhing, hot plate, and Von Frey tests. Additionally, the study investigated the modulation of various pain signaling pathways using a pharmacological approach. Results: The results demonstrated that all doses of the HEXA significantly increased latency in the hot plate test, decreased the number of abdominal contortions, reduced hyperalgesia in the Von Frey test, and reduced both phases of the formalin test. Caffeic acid reduced licking time in the first phase of the formalin test at all doses, with the highest dose showing significant effects in the second phase. The HEXA potentially modulated α₂-adrenergic (52.99%), nitric oxide (57.77%), glutamatergic (33.66%), vanilloid (39.84%), cyclic guanosine monophosphate (56.11%), and K⁺ATP channel-dependent pathways (38.70%). Conversely, CA influenced the opioid, glutamatergic (53.60%), and vanilloid (34.42%) pathways while inhibiting nitric oxide (52.99%) and cyclic guanosine monophosphate (38.98%). Conclusions: HEXA and CA exhibit significant antinociceptive effects due to their potential interference in multiple pain signaling pathways. While the molecular targets remain to be fully investigated, HEXA and CA demonstrate significant potential for the development of new analgesic drugs. Full article