Nowadays, the need to efficiently process information in Internet of Things (IoT)-oriented heterogeneous scenarios has increased significantly, e.g., in all scenarios where unobtrusive environmental monitoring is bene...
详细信息
ISBN:
(纸本)9798350304572
Nowadays, the need to efficiently process information in Internet of Things (IoT)-oriented heterogeneous scenarios has increased significantly, e.g., in all scenarios where unobtrusive environmental monitoring is beneficial for the involved people (e.g., inside public transport vehicles, indoor workplaces and offices, large public infrastructures, etc.). This objective typically requires the combination of heterogeneous IoT systems, which need to efficiently share information, e.g., through the Web of Things (WoT) paradigm. In this paper, we propose an edge computing-oriented flexible WoT architecture, with distributed intelligence, for air quality monitoring and prediction inside a public transport bus. Our results show that the proposed architecture allows seamless integration of heterogeneous IoT systems according to a WoT perspective, exploiting the device/edge/fog computing continuum and using containerized and secure processing modules.
The proceedings contain 154 papers. The topics discussed include: transient stability analysis of power system considering multiple inverter aggregation model;development of the safeguard companion for secondary safet...
ISBN:
(纸本)9798331533694
The proceedings contain 154 papers. The topics discussed include: transient stability analysis of power system considering multiple inverter aggregation model;development of the safeguard companion for secondary safety measures of relay protection;an open interconnection system for computing power based on service mesh;securing distributed power dispatching systems: a framework for cybersecurity and 5G integration;utilizing deep belief networks for power system state estimation and anomaly detection;fire detection approach for 10kV distribution room based on spatial perception and hybrid attention mechanism;research on the application of information sharing technology in distributed photovoltaic aggregation optimization control strategy;and construction and optimization of key quality problem graph of power equipment based on association rule mining.
distributed databases are often used when scalability, fault tolerance, and high availability are crucial. They excel in scenarios where traditional, centralized databases may struggle to handle the increasing volume ...
详细信息
This paper explores the novel use of ferroelectric field-effect transistors (FeFETs) in a mixed multi-level cell (MLC) and single-level cell (SLC) configuration, aiming to strike an optimal balance between area effici...
详细信息
ISBN:
(纸本)9798350383638;9798350383645
This paper explores the novel use of ferroelectric field-effect transistors (FeFETs) in a mixed multi-level cell (MLC) and single-level cell (SLC) configuration, aiming to strike an optimal balance between area efficiency and data integrity. Addressing the inherent trade-offs in MLC configurations, which offer high bit capacity at the cost of increased error rates, we propose a mixed mapping scheme that effectively combines the advantages of both MLC and SLC configurations. Further, this study introduces a specialized fitting algorithm designed to identify the optimal configuration. Complemented by simulated annealing for hyperparameter tuning, this approach not only proves efficacious in this specific context but also offers adaptability for diverse design objectives.
In distributedcomputing environments, computation offloading is a vital strategy for maximizing the performance and energy efficiency of mobile devices. distributed deep learning-based offloading (DDLO) [10] and deep...
详细信息
The rapid evolution of mobile networks presents challenges for devices with limited computing power. Mobile or multi-access edge computing (MEC) addresses this by providing computing resources in proximity to end devi...
详细信息
The convergence of SGD based distributed training algorithms is tied to the data distribution across workers. Standard partitioning techniques try to achieve equal-sized partitions with per-class population distributi...
详细信息
ISBN:
(纸本)9798350307443
The convergence of SGD based distributed training algorithms is tied to the data distribution across workers. Standard partitioning techniques try to achieve equal-sized partitions with per-class population distribution in proportion to the total dataset. Partitions having the same overall population size or even the same number of samples per class may still have Non-IID distribution in the feature space. In heterogeneous computing environments, when devices have different computing capabilities, even-sized partitions across devices can lead to the straggler problem in distributed SGD. We develop a framework for distributed SGD in heterogeneous environments based on a novel data partitioning algorithm involving submodular optimization. Our data partitioning algorithm explicitly accounts for resource heterogeneity across workers while achieving similar class-level feature distribution and maintaining class balance. Based on this algorithm, we develop a distributed SGD framework that can accelerate existing SOTA distributed training algorithms by up to 32%.
Deep learning hardware accelerators commonly incorporate a substantial quantity of multiplier units. Yet, the considerable complexity of multiplier circuits renders them a bottleneck, contributing to increased costs a...
详细信息
ISBN:
(纸本)9798350383638;9798350383645
Deep learning hardware accelerators commonly incorporate a substantial quantity of multiplier units. Yet, the considerable complexity of multiplier circuits renders them a bottleneck, contributing to increased costs and latency. Approximate computing proves to be an effective strategy for mitigating the overhead associated with multipliers. This paper introduces an original approximation technique for signed multiplication on FPGAs. The approach involves a novel segmentation method applied to the Baugh-Wooley multiplication algorithm. Each segment is optimally accommodated within look-up table resources of modern AMD-Xilinx FPGA families. The paper details the design of an INT8 multiplier using the proposed approach, presenting implementation results and accuracy assessments for the inference of benchmark deep learning models. The implementation results reveal significant savings of 53.6% in LUT utilization compared to the standard INT8 Xilinx multiplier. Accuracy measurements conducted on four popular deep learning benchmarks show an average accuracy degradation of 4.8% in post-training deployment and 0.7% after retraining. The source code for this work is available on GitHub(1).
In the booming field of quantum computing, Grover’s Algorithm emerges as a pivotal quantum search algorithm, theoretically capable of outperforming classical brute-force search methods by exploiting the principles of...
详细信息
The complexity inherent in managing cloud computingsystems calls for novel solutions that can effectively enforce high-level Service Level Objectives (SLOs) promptly. Unfortunately, most of the current SLO management...
详细信息
ISBN:
(纸本)9798350304817
The complexity inherent in managing cloud computingsystems calls for novel solutions that can effectively enforce high-level Service Level Objectives (SLOs) promptly. Unfortunately, most of the current SLO management solutions rely on reactive approaches, i.e., correcting SLO violations only after they have occurred. Further, the few methods that explore predictive techniques to prevent SLO violations focus solely on forecasting low-level system metrics, such as CPU and Memory utilization. Although valid in some cases, these metrics do not necessarily provide clear and actionable insights into application behavior. This paper presents a novel approach that directly predicts high-level SLOs using low-level system metrics. We target this goal by training and optimizing two state-of-the-art neural network models, a Short-Term Long Memory LSTM-, and a Transformer-based model. Our models provide actionable insights into application behavior by establishing proper connections between the evolution of low-level workload-related metrics and the high-level SLOs. We demonstrate our approach to selecting and preparing the data. We show in practice how to optimize LSTM and Transformer by targeting efficiency as a high-level SLO metric and performing a comparative analysis. We show how these models behave when the input workloads come from different distributions. Consequently, we demonstrate their ability to generalize in heterogeneous systems. Finally, we operationalize our two models by integrating them into the Polaris framework we have been developing to enable a performance-driven SLO-native approach to Cloud computing.
暂无评论