Latency Reduction Research Articles

Blockchain-Based Caching Architecture for DApp Data Security and Delivery.

Decentralized applications (DApps) built on blockchain technology offer a promising solution to issues caused by centralization. However, traditional DApps leveraging off-chain storage face performance challenges due to factors such as storage location, network speed, and hardware conditions. For example, decentralized storage solutions such as IPFS suffer from diminished download performance due to I/O constraints influenced by data access patterns. Aiming to enhance the Quality of Service (QoS) in DApps built on blockchain technology, this paper proposes a blockchain node-based distributed caching architecture that guarantees real-time responsiveness for users. The proposed architecture ensures data integrity and user data ownership through blockchain while maintaining cache data consistency through local blockchain data. By implementing local cache clusters on blockchain nodes, our system achieves rapid response times. Additionally, attribute-based encryption is applied to stored content, enabling secure content sharing and access control, which prevents data leakage and unauthorized access in unreliable off-chain storage environments. Comparative analysis shows that our proposed system achieves a reduction in request processing latency of over 89% compared to existing off-chain solutions, maintaining cache data consistency and achieving response times within 65 ms. This demonstrates the model's effectiveness in providing secure and high-performance DApp solutions.

Mechanism of electroacupuncture for vascular dementia based on proteomics

To explore the potential mechanism of electroacupuncture (EA) for vascular dementia (VD) using tandem mass tag (TMT) quantitative proteomics technology. Among 80 male SPF SD rats, 78 rats which met the selection criteria through the Morris water maze test were selected and randomly divided into a sham surgery group (18 rats) and a surgery group (60 rats). VD model was established by four-vessel occlusion (4-VO) method in the surgery group, and 36 rats with successful modeling were randomly assigned to a model group (18 rats) and an EA group (18 rats). Each group was further divided into three subgroups based on intervention duration, with each subgroup containing 6 rats. Seven days after model establishment, the EA group received EA intervention at left and right "Sishencong" (EX-HN 1) and bilateral "Fengchi" (GB 20), with continuous wave at a frequency of 2 Hz and current intensity of 1 mA, daily for 30 min, with subgroups receiving EA for 7, 14, or 21 d respectively. Cognitive function before and after interventions was assessed using Morris water maze. Proteomic analysis was conducted on the optimal EA subgroup and corresponding sham surgery and model subgroups, identifying differentially expressed proteins and analyzing them through bioinformatics. Differentially expressed target proteins was performed using parallel reaction monitoring (PRM) and Western blot techniques. Compared to the sham surgery group, the model group exhibited prolonged escape latency and reduced number of platform crossings (P<0.01); compared with model group, the EA group showed reductions in escape latency and increased platform crossings after 7, 14, and 21 days of intervention (P<0.01, P<0.05). Compared to the 7 and 14-day intervention, the rats in the EA group of 21-day intervention showed the most significant improvements in reductions of escape latency and increased platform crossings (P<0.01, P<0.05), and was selected for further proteomic, PRM analyses, and Western blot validation. Compared to the sham surgery group, the model group displayed 71 differentially expressed proteins, with 50 up-regulated and 21 down-regulated proteins; compared to the model group, the EA group had 54 differentially expressed proteins, with 30 up-regulated and 24 down-regulated proteins. Functional enrichment and clustering analyses indicated that these proteins were primarily associated with cellular processes, metabolic processes, phagocytosis recognition, immune response, and regulation of extracellular matrix, etc. Enrichment was observed in the mammalian target of rapamycin (mTOR) signaling pathway and neurotrophic factors signaling pathways, involving glycogen synthase kinase 3β (GSK3β) and mitogen-activated protein kinase kinase 2 (Map2k2), with PRM and Western blot findings consistent with the proteomic results. Which meant that compared with the model group, the protein expression of GSK3β and Map2k2 of hippocampus was increased in the EA group (P<0.01, P<0.05). EA at "Sishencong" (EX-HN 1) and "Fengchi" (GB 20) could improve cognitive function in VD rats, with the mechanism involving multiple targets and pathways, potentially related to GSK3β, Map2k2 proteins, and the mTOR and neurotrophic factor signaling pathways.

Modeling Fragile X Syndrome: Characterizing Fmr1 Gene Knockout Mice across Genotype, Behavior, and Morphology

ABSTRACT Objective: Fragile X syndrome (FXS) is a hereditary condition resulting from dynamic mutations in the Fmr1 gene, leading to reduced or absent fragile X mental retardation protein (FMRP). Although molecular genetic diagnostics for FXS have advanced, there is currently a lack of effective clinical treatment. Hence, identifying a suitable FXS animal model is crucial for FXS therapeutic studies. This study primarily aims to evaluate whether Fmr1 mice display aberrant behaviors similar to those observed in human FXS patients, aiming to validate their feasibility as a model for FXS. Research Design and Methods: Polymerase chain reaction and immunoblotting were employed to genotype Fmr1 knockout (KO) mice and wild-type (WT) mice. Using hematoxylin and eosin staining and Golgi staining, testicular and dendritic spine morphology in both animal groups was observed. The animals were subjected to audiogenic seizure (AGS) observation, evaluating their learning and memory levels using the water maze and light avoidance methods, while their exploratory behavior capacity was assessed using the open-field and autonomous activity experiments. Anxiety behavior was evaluated using the elevated plus maze. Results: WT mice produced a 500 bp DNA fragment, while KO mice yielded an 800 bp fragment. KO mice exhibited a lack of FMRP expression, notably increased dendritic spine length and density, and presented with macroorchidism. KO mice showed increased susceptibility to AGS. In the Morris water maze experiment, KO mice displayed significantly increased escape latency and platform crossings. The spatial search experiment revealed that KO mice spent significantly less time in the target quadrant compared to WT mice. The light-avoidance and platform experiments indicated a significant reduction in latency for KO mice but an increase in errors. In the open-field experiment, KO mice spent significantly more time and made more entries into the central zone compared to WT mice. KO mice exhibited increased total travel distance and speed. The elevated plus maze experiment showed that KO mice spent significantly more time, made more entries, and traveled further in the open area compared to WT mice. The autonomous activity experiment revealed that KO mice had significantly more activity counts than WT mice but fewer standing counts. Conclusions: Findings from genotype, protein expression spectrum, morphology, and behavioral results suggest that Fmr1 gene KO mice demonstrate phenotypes similar to those observed in human FXS, making them a potential model for future FXS research. These phenotypes are associated with the absence of FMRP expression.

Enhanced Wireless Communication Optimization with Neural Networks, Proximal Policy Optimization and Edge Computing for Latency and Energy Efficiency

This research proposes a novel approach for efficient resource allocation in wireless communication systems. It combines dynamic neural networks, Proximal Policy Optimization (PPO), and Edge Computing Orchestrator (ECO) for latency-aware and energy-efficient resource allocation. The proposed system integrates multiple components, including a dynamic neural network, PPO, ECO, and a Mobile Edge Computing (MEC) server. The experimental methodology involves utilizing the NS-3 simulation platform to assess latency and energy efficiency in resource allocation within a wireless communication network, incorporating an ECO, MEC server, and dynamic task scheduling algorithms. It demonstrates a holistic and adaptable approach to resource allocation in dynamic environments, showcasing a notable reduction in latency for devices and tasks. Latency values range from 5 to 20 milliseconds, with corresponding resource utilization percentages varying between 80% and 95%. Additionally, energy-efficient resource allocation demonstrates a commendable reduction in energy consumption, with measured values ranging from 10 to 30 watts, coupled with efficient resource usage percentages ranging from 70% to 85%. These outcomes validate the efficacy of achieving both latency-aware and energy-efficient resource allocation for enhanced wireless communication systems. The proposed system has broad applications in healthcare, smart cities, IoT, real-time analytics, autonomous vehicles, and augmented reality, offering a valuable solution to optimize energy consumption, reduce latency, and enhance system efficiency in these industries.

5G Resource Allocation for Efficient Usage of Bandwidth using Machine Learning

Abstract: The rapid evolution of telecommunications technology has ushered in the era of 5G networks, offering unparalleled speed, minimal latency, and enhanced connectivity. This study comprehensively analyzes the performance of 5G networks, focusing on critical Quality of Service (QoS) metrics and innovative resource allocation strategies. Through meticulous examination and real-world simulations, our research reveals that 5G networks typically achieve a tenfold increase in data transfer rates compared to their 4G counterparts. Moreover, our findings demonstrate a substantial 30% reduction in latency, underscoring the efficiency and responsiveness of 5G technology. Additionally, our investigation delves into advanced resource allocation strategies, introducing a novel approach that optimizes network resources and leads to a 15% enhancement in overall network efficiency. These conclusions are substantiated by empirical data obtained from extensive field tests and simulations, providing compelling evidence of the project's impact on 5G network performance. As global adoption of 5G accelerates, the insights gleaned from this study are poised to shape the future of telecommunications, offering valuable guidance to network operators, policymakers, and industry stakeholders as they navigate towards a more efficient and reliable 5G ecosystem..

Quad-functioning parity layout for nanocomputing: A QCA design

Quantum dot cellular automata (QCA) is considered an alternative to conventional technologies like CMOS (Complementary Metal-Oxide-Semiconductor) technology due to its potential for lower power consumption, higher speed, and increased device density. QCA introduces a novel approach to designing nano communication circuits and systems. Nano communications data mistakes are detected via parity generators and checkers. The parity bit of each data block ensures that the number of 1′s is either even or odd. Consequently, the system requires four circuits: an even parity generator, an odd parity generator, an even parity checker, and an odd parity checker. The whole system requires more space and cell complexity. In this work, we propose a QCA architecture that serves as a generator for both even and odd parities, as well as a checker for both even and odd parities. It is a quad-functioning circuit that performs four distinct operations within a single design, utilizing 118 QCA cells and occupying an area of 0.17 μm2. The recommended approach uses an efficient XOR gate, resulting in improvements across several performance metrics. QCAPro calculates energy dissipation and design parameters. The recommended QCA circuit outperformed similar QCA circuits in size, complexity, and energy dissipation. The circuit's design cost functions are also low. There has been a 17% reduction in latency and an 86% improvement in QCA-specific costs when compared to the optimal existing design. Moreover, it necessitates a 40% reduction in majority gate usage. The proposed design may compete effectively with other equivalent higher-order circuit designs by reducing the need for multiple blocks in conventional circuits to execute the same task. This architecture holds potential benefits for nano processors and nano communication networks.

Does responsibility for potential harm attenuate the effects of repeated checking on metamemory and automatization?

Background and objectivesRepeated checking results in large reductions in metamemory variables (confidence, details, and vividness). It has been suggested that the underlying mechanism is gradual automatization. At the same time, individuals with obsessive-compulsive disorder (OCD) are reluctant to automatize routine processes. The aim was to investigate whether high responsibility for potential harm, typical of OCD, would attenuate the effects of repeated checking on metamemory variables and automatization. MethodsOne hundred seventy-five participants were initially provided with a cover story that put the subsequent virtual checking task in a context of potential harm for not checking properly. Participants were randomly allocated to four experimental groups (varying high and low responsibility, relevant and irrelevant checking) and performed a virtual checking task repeatedly, using either identical stimuli (relevant checking) or different stimuli (irrelevant checking) between the first and final checking trial. Metamemory variables were rated on visual analogue scales, and response latencies were assessed to establish automatization. ResultsLarger reductions in metamemory variables following relevant checking compared to irrelevant checking replicated previous findings. High responsibility did not affect these results. Large reductions in response latencies across the checking trials (automatization) were also independent of the perceived responsibility. LimitationsWe did not include individuals with OCD. ConclusionsSince responsibility did not influence the effects of repeated checking on metamemory variables, findings are consistent with the idea that automatization remains a plausible explanation of the effects of repeated checking on metamemory variables in individuals with OCD.

Deep Learning Model Compression Techniques Performance on Edge Devices

Artificial intelligence at the edge can help solve complex tasks faced by various sectors such as automotive, healthcare and surveillance. However, challenged by the lack of computational power from the edge devices, artificial intelligence models are forced to adapt. Many have developed and quantified model compres-sion approaches over the years to tackle this problem. However, not many have considered the overhead of on-device model compression, even though model compression can take a considerable amount of time. With the added metric, we provide a more complete view on the efficiency of model compression on the edge. The objective of this research is identifying the benefit of compression methods and it’s tradeoff between size and latency reduction versus the accuracy loss as well as compression time in edge devices. In this work, quantitative method is used to analyze and rank three common ways of model compression: post-training quantization, unstructured pruning and knowledge distillation on the basis of accuracy, latency, model size and time to compress overhead. We concluded that knowledge distillation is the best, with potential of up to 11.4x model size reduction, and 78.67% latency speed up, with moderate loss of accura-cy and compression time.

A generic deep learning architecture optimization method for edge device based on start-up latency reduction

In the promising Artificial Intelligence of Things technology, deep learning algorithms are implemented on edge devices to process data locally. However, high-performance deep learning algorithms are accompanied by increased computation and parameter storage costs, leading to difficulties in implementing huge deep learning algorithms on memory and power constrained edge devices, such as smartphones and drones. Thus various compression methods are proposed, such as channel pruning. According to the analysis of low-level operations on edge devices, existing channel pruning methods have limited effect on latency optimization. Due to data processing operations, the pruned residual blocks still result in significant latency, which hinders real-time processing of CNNs on edge devices. Hence, we propose a generic deep learning architecture optimization method to achieve further acceleration on edge devices. The network is optimized in two stages, Global Constraint and Start-up Latency Reduction, and pruning of both channels and residual blocks is achieved. Optimized networks are evaluated on desktop CPU, FPGA, ARM CPU, and PULP platforms. The experimental results show that the latency is reduced by up to 70.40%, which is 13.63% higher than only applying channel pruning and achieving real-time processing in the edge device.

Energy-efficient buildings with energy-efficient optimized models: a case study on thermal bridge detection

Thermographic inspection is particularly effective in identifying thermal bridges because it visualizes temperature differences on the building’s surface. The focus of this work is on energy-efficient computing for deep learning-based thermal bridge (anomaly) detection models. In this study, we concentrate on object detection-based models such as Mask R-CNN_FPN_50, Swin-T Transformer, and FSAF. We do benchmark tests on TBRR dataset with varying input sizes. To overcome the energy-efficient design, we apply optimizations such as compression, latency reduction, and pruning to these models. After our proposed improvements, the inference of the anomaly detection model, Mask R-CNN_FPN_50 with compression technique, is approximately 7.5% faster than the original. Also, more acceleration is observed in all models with increasing input size. Another criterion we focus on is total energy gain for optimized models. Swin-T transformer has the most inference energy gains for all input sizes (≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document}27 J for 3000 x 4000 and ≈\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\approx$$\\end{document}14 J for 2400 x 3400). In conclusion, our study presents an optimization of size, weight, and power for vision-based anomaly detection for buildings.

Does Transcranial Direct Current Stimulation Affect Potential P300-Related Events in Vascular Dementia? Considerations from a Pilot Study.

Vascular dementia, the second most common type of dementia, currently lacks a definitive cure. In the pursuit of therapies aimed at slowing its progression and alleviating symptoms, transcranial direct current stimulation (tDCS) emerges as a promising approach, characterized by its non-invasive nature and the ability to promote brain plasticity. In this study, the primary objective was to investigate the effects of a two-week cycle of tDCS on the dorsolateral prefrontal cortex (DLPFC) and neurophysiological functioning in thirty patients diagnosed with vascular dementia. Each participant was assigned to one of two groups: the experimental group, which received anodal tDCS to stimulate DPCFL, and the control group, which received sham tDCS. Neurophysiological functions were assessed before and after tDCS using P300 event-related potentials (ERPs), while neuropsychological function was evaluated through a Mini-Mental State Examination (MMSE). The results showed a reduction in P300 latency, indicating a faster cognitive process; an increase in P300 amplitude, suggesting a stronger neural response to cognitive stimuli; and a significant improvement in MMSE scores compared to the control group, indicating an overall enhancement in cognitive functions. These findings suggest that tDCS could represent a promising therapeutic option for improving both neurophysiological and cognitive aspects in patients with vascular dementia.

NIR-II light therapy improves cognitive performance in MPTP induced Parkinson's disease rat models: A preliminary experimental study

Cognitive impairment is an important component of non motor symptoms in Parkinson's disease (PD), and if not addressed in a timely manner, it can easily progress to dementia. However, no effective method currently exists to completely prevent or reverse cognitive impairment associated with PD. We therefore aimed to investigate the therapeutic effect of near-infrared region II light (NIR-II) region illumination on cognitive impairment in PD through behavioral experiments (water maze and rotary rod) and multiple fluorescence immunohistochemistry techniques. The 1-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridine (MPTP)-induced group was compared with the MPTP- untreated rat group, showing a significant reduction in escape latency and significant increase in the fall latency in the MPTP-treated group. The horizontal analysis results indicated that NIR-II phototherapy improved the learning and cognitive abilities as well as coordination and balance abilities of rats. Post-treatment, the MPTP rats showed significantly shortened, escape latency, prolonged target quadrant residence time, and prolonged fall latency compared with pre-treatment. The longitudinal analysis results reaffirmed that NIR-II phototherapy improved the learning and cognitive abilities as well as coordination and balance abilities of rats. The multiple fluorescence immunohistochemistry analysis trend plot showed that the activated microglia and astrocytes in the hippocampus were highest in MPTP-induced PD untreated group, moderate in MPTP-induced PD treatment group, and lowest in the control group. Our data indicates that NIR-II illumination improves learning and cognitive impairment as well as coordination and balance abilities in PD rats by downregulating the activation of microglia and astrocytes in the hippocampus.

Incremental Sliding Window Connectivity over Streaming Graphs

We study index-based processing for connectivity queries within sliding windows on streaming graphs. These queries, which determine whether two vertices belong to the same connected component, are fundamental operations in real-time graph data processing and demand high throughput and low latency. While indexing methods that leverage data structures for fully dynamic connectivity can facilitate efficient query processing, they encounter significant challenges with deleting expired edges from the window during window updates. We introduce a novel indexing approach that eliminates the need for physically performing edge deletions. This is achieved through a unique bidirectional incremental computation framework, referred to as the BIC model. The BIC model implements two distinct incremental computations to compute connected components within the window, operating along and against the timeline, respectively. These computations are then merged to efficiently compute queries in the window. We propose techniques for optimized index storage, incremental index updates, and efficient query processing to improve BIC effectiveness. Empirically, BIC achieves a 14× increase in throughput and a reduction in P95 latency by up to 3900× when compared to state-of-the-art indexes.

Reducing False Alarms in Wearable Seizure Detection With EEGformer: A Compact Transformer Model for MCUs.

The long-term, continuous analysis of electroencephalography (EEG) signals on wearable devices to automatically detect seizures in epileptic patients is a high-potential application field for deep neural networks, and specifically for transformers, which are highly suited for end-to-end time series processing without handcrafted feature extraction. In this work, we propose a small-scale transformer detector, the EEGformer, compatible with unobtrusive acquisition setups that use only the temporal channels. EEGformer is the result of a hardware-oriented design exploration, aiming for efficient execution on tiny low-power micro-controller units (MCUs) and low latency and false alarm rate to increase patient and caregiver acceptance.Tests conducted on the CHB-MIT dataset show a 20% reduction of the onset detection latency with respect to the state-of-the-art model for temporal acquisition, with a competitive 73% seizure detection probability and 0.15 false-positive-per-hour (FP/h). Further investigations on a novel and challenging scalp EEG dataset result in the successful detection of 88% of the annotated seizure events, with 0.45 FP/h.We evaluate the deployment of the EEGformer on three commercial low-power computing platforms: the single-core Apollo4 MCU and the GAP8 and GAP9 parallel MCUs. The most efficient implementation (on GAP9) results in as low as 13.7 ms and 0.31 mJ per inference, demonstrating the feasibility of deploying the EEGformer on wearable seizure detection systems with reduced channel count and multi-day battery duration.

From land to water: “Sunken” T-maze for associated learning in cichlid fish

The study introduced and evaluated learning paradigms for Maylandia callainos cichlids using a modified version of the rodent T-maze, filled with tank water (the "sunken" modification). Both male and female fish underwent training in two distinct conditioning paradigms. Firstly, simple operant conditioning involved placing a food reward in either the right or left compartment. Cichlids demonstrated the ability to purposefully find the bait within 6 days of training, with a persistent place preference lasting up to 6 days. Additionally, the learning dynamics varied with sex: female cichlids exhibited reduction in latency to visit the target compartment and consume the bait, along with a decrease in the number of errors 3 and 4 days earlier than males, respectively. Secondly, visually-cued operant conditioning was conducted, with a food reward exclusively placed in the yellow compartment, randomly positioned on the left or right side of the maze during each training session. Visual learning persisted for 10 days until reaction time improvement plateaued. Color preference disappeared after 4 consecutive check-ups, with no sex-related interference. For further validation of visually-cued operant conditioning paradigm, drugs MK-801 (dizocilpine) and caffeine, known to affect performance in learning tasks, were administered intraperitoneally. Chronic MK-801 (0.17 mg/kg) impaired maze learning, resulting in no color preference development. Conversely, caffeine administration enhanced test performance, increasing precision in fish. This developed paradigm offers a viable approach for studying learning and memory and presents an effective alternative to rodent-based drug screening tools, exhibiting good face and predictive validity.

Mobility-aware proactive video caching based on asynchronous federated learning in mobile edge computing systems

With the rapid advance of the Internet of Things (IoT) and the surging user-centric smart devices, video services such as short videos, real-time video streaming, and Virtual Reality (VR) gaming are emerging. Traditional cloud-centric caching is no longer able to satisfy the low-latency needs of mobile user groups. Predicting the most popular content in advance and caching it on edge devices such as small base stations (SBS) is a promising solution to alleviate network congestion and supply better Quality of Service (QoS) for mobile users. However, frequently changing popular content, limited edge caching resources, privacy of user data, and user mobility pose significant challenges to edge video caching. To tackle these challenges, we propose an efficient asynchronous federated learning (AFL)-based mobility-aware proactive video caching scheme (AF-MPVC), which significantly improves the overall network performance while preventing the risk of user data leakage. First, we model the proactive video caching problem in mobile edge computing systems and consider the impact of user mobility on AFL, representing the problem as an autoencoder (AE) model loss minimization problem. Then, a proactive video caching algorithm is proposed based on a filtering model, which uses the hidden features obtained from the trained AE model to compute the user similarity to find the popular videos that are most likely to be accessed by users and combine them with the most frequently requested videos. Meanwhile, we take into account the short duration of high mobility users staying on the current SBS and cache some videos to neighboring edge servers to have more redundant space to store more less popular videos. Finally, experiments on three real-world datasets demonstrate that the presented caching scheme outperforms other baseline caching schemes with respect to cache hit rate and latency reduction.

Effects of different mylohyoid muscle stimulations on swallowing cortex excitability in healthy subjects

BackgroundDysphagia has been recognized by the World Health Organization as a medical disability. Improving mylohyoid muscle function plays an important role in pharyngeal dysphagia. The aim of this study was to evaluate the treatment of transcranial magnetic stimulation (TMS), peripheral magnetic stimulation (PMS), and electrical stimulation (ES) for dysphagia. MethodsForty healthy subjects were randomly divided into four groups: TMS+PMS, TMS, PMS, and ES. TMS stimulated the cortical representative area of the mylohyoid muscle and the PMS was directly stimulating the mylohyoid muscle, both of them at a frequency of 10 Hz for a total of 1,800 pulses. The intensity of ES was based on the subject’s tolerance level, usually 2–5 mA. Functional near infrared spectroscopy (fNIRS) and motor evoked potential (MEP) of the mylohyoid muscle were used to evaluate the immediate effects of stimulation on swallowing cortex excitability of healthy subjects before and after intervention. ResultsThe fNIRS results revealed notable activation across multiple channels in the four groups of healthy subjects both pre- and post- the intervention. Among these channels, the activation levels were most pronounced in the TMS+PMS group, followed by the TMS, PMS, and ES groups, respectively. Regarding the MEP results, post-intervention observations indicated a reduction in bilateral latency and an increase in bilateral amplitude in the TMS+PMS group. Additionally, the left amplitude exhibited an increase in the TMS group. ConclusionsIn fNIRS, all four stimulation methods significantly activated the swallowing cortex of healthy subjects, and the activation of TMS+PMS was the most obvious, followed by TMS, PMS, and ES.

Smart Offloading Strategies by Optimizing Fog Computing Through Reinforcement Learning Strategies

With the proliferation of Internet of Things(IoT) devices and the exponential growth in data volume, fog computing has emerged as a promising paradigm to address the limitations of cloud-centric architectures by bringing computation and storage closer to the data source. In fog comp uting environments, efficient task offloading plays a crucial role in optimizing resource utilization and minimizing latency. This paper proposes a novel approach to smart offloading strategies utilizing reinforcement learning (RL) techniques. To validate the effectiveness of our approach, we conduct extensive simulations and experiments using representative fog computing scenarios. Results demonstrate significant improvements in offloading efficiency, latency reduction, and resource utilization compared to traditional static offloading methods. Keywords: Fog Computing, Internet of Things (IoT), offloading, reinforcement learning.

Real-time fusion multi-tier DNN-based collaborative IDPS with complementary features for secure UAV-enabled 6G networks

UAV-enabled Integrated Sensing and Communication (ISAC) in sixth-generation (6G) wireless networks has sparked significant research interest. UAVs are positioned as aerial wireless platforms, extending coverage and improving Sensing and Communication (S&C) services. However, integrating UAVs introduces vulnerabilities due to multi-connectivity, necessitating robust security measures. To address this, efforts are focused on developing effective Intrusion Detection Systems (IDS). Here ML-based IDSs depend on the most suited Machine Learning (ML) algorithms for improved detection accuracy. However, inadequate detection features frequently contribute to the limitations of detection accuracy in various emerging cyber-attacks. Our study explores attack traits in UAV-enabled 6G networks, using complementary detection features to enhance ML-based attack detection. Additionally, existing Deep Neural Network (DNN) solutions for UAVs-enabled 6G networks mainly use single models (e.g., CNN, SVM, CGAN) instead of multi-tier models, (e.g., Hybrid, Fusion, Ensemble). Here, although the use of a single model can be faster for computer systems, it is difficult to understand the growing complexity of intrusion patterns in data. Also, these single models might not be good at recognizing the unique patterns of less common issues in datasets. Thus, we propose a fusion multi-tier DNN-based Collaborative Intrusion Detection and Prevention System (CIDPS) for critical UAV-enabled 6G networks. This system boosts accuracy without sacrificing latency reductions. Our approach learns decision boundaries from imbalanced data points using preceding DNNs sequentially. Moreover, most existing research works do not focus on intrusion prevention methods and deployment frameworks for UAV networks. This research proposes effective IPS mechanisms and deployment architecture for CIDPS in UAV-enabled 6G networks. It incorporates a CIDPS with an emergency response protocol, neutralizing attacks upon anomaly detection. We validate our solution through diverse dataset experiments (UAVIDS-2020, NF-UQ-NIDS-v2, 5G-NIDD). Unlike traditional practices where IDS are simulated, we implement CIDPS on actual UAV devices (PX4 Vision Dev Kit V1.5, DJI mini se, DJI mini 3 pro) and real-world UAV networks. Our approach outperforms prior algorithms with a 99.25% attack classification accuracy, higher detection efficiency, and lower resource usage, exceeding 99.05% detection rates.

Reg-Tune: A Regression-Focused Fine-Tuning Approach for Profiling Low Energy Consumption and Latency

Fine-tuning deep neural networks is pivotal for creating inference modules that can be suitably imported to edge or field-programmable gate array (FPGA) platforms. Traditionally, exploration of different parameters throughout the layers of deep neural networks has been done using grid search and other brute force techniques. Although these methods lead to the optimal choice of network parameters, the search process can be very time consuming and may not consider deployment constraints across different target platforms. This work addresses this problem by proposing Reg-Tune, a regression-based profiling approach to quickly determine the trend of different metrics in relation to hardware deployment of neural networks on tinyML platforms like FPGAs and edge devices. We start by training a handful of configurations belonging to different combinations of \(\mathcal {NN}\scriptstyle \langle q (quantization),\,s (scaling)\rangle \displaystyle\) or \(\mathcal {NN}\scriptstyle \langle r (resolution),\,s\rangle \displaystyle\) workloads to generate the accuracy values respectively for their corresponding application. Next, we deploy these configurations on the target device to generate energy/latency values. According to our hypothesis, the most energy-efficient configuration suitable for deployment on the target device is a function of the variables q , r , and s . Finally, these trained and deployed configurations and their related results are used as data points for polynomial regression with the variables q , r , and s to realize the trend for accuracy/energy/latency on the target device. Our setup allows us to choose the near-optimal energy-consuming or latency-driven configuration for the desired accuracy from the contour profiles of energy/latency across different tinyML device platforms. To this extent, we demonstrate the profiling process for three different case studies and across two platforms for energy and latency fine-tuning. Our approach results in at least 5.7 \(\times\) better energy efficiency when compared to recent implementations for human activity recognition on FPGA and 74.6% reduction in latency for semantic segmentation of aerial imagery on edge devices compared to baseline deployments.