To enable the large scale and efficient deployment of Artificial Intelligence (AI), confluence AI Edge Computing has given rise to Intelligence, which leverages on computation communication capabilities end devices edge servers process data closer where it is produced. One enabling technologies privacy preserving machine learning paradigm known as Federated Learning (FL), enables owners conduct model training without having transmit their raw third-party servers. However, FL network...

Amid growing concerns on data privacy, Federated Learning (FL) has emerged as a promising privacy preserving distributed machine learning paradigm. Given that the FL network is expected to be implemented at scale, several studies have proposed system architectures towards improving scalability and efficiency. Specifically, Hierarchical (HFL) utilizes cluster heads, e.g., base stations, for intermediate aggregation relay of model parameters. Serverless also recently, in which owners, i.e.,...

For future Internet of Things (IoT) systems, data-driven and dynamic spectrum-sharing schemes can significantly improve the spectrum utilization efficiency. However, conventional centralized architecture such IoT systems is often considered to be nontransparent, costly, vulnerable potential attacks single-point failures. To address aforementioned issues, a blockchain-based scheme has been proposed investigated in this work, which aims at enhancing system by providing desirable features, as...

Internet of Things (IoT) technology enables various physical devices to collect, process and exchange information. Market oriented models become important for IoT systems efficiently utilize information, as network nodes operate in a highly distributed autonomous manner. In this work, we propose three-player game theoretic market model information trading, considering direct indirect externalities among participants. the model, an service provider collects processes then delivers processed...

Deep learning has revolutionized computer vision and other fields since its big bang in 2012. However, it is challenging to deploy Neural Networks (DNNs) into real-world applications due their high computational complexity. Binary (BNNs) dramatically reduce complexity by replacing most arithmetic operations with bitwise operations. Existing implementations of BNNs have been focusing on GPU or FPGA, using the conventional image-to-column method that doesn't perform well for binary convolution...

In cloud computing, services can be allocated to users upon requests in an on-demand basis. Heterogeneous service providers may join the systems serve various types of users. Cloud complementary or substitutable. For services, request for a bundle e.g., CPU and storage, gain higher benefit from requesting them alone. The substitutable have similar functionalities users, different database obtaining one replace another one. Furthermore, also influence each other because externalities,...

Simulations provide a flexible and valuable method to study the behaviors of information propagation over complex social networks. High Performance Computing (HPC) is technology that allows implementation efficient algorithms on powerful new hardware resources. With increased computing resource usage in large-scale network based simulations, it therefore attractive apply emerging HPC techniques improve simulation performance. This paper describes optimized strategies algorithmic adaptation...

Distributed artificial intelligence (AI) is becoming an efficient approach to fulfill the high and diverse requirements for future vehicular networks. However, distributed tasks generated by vehicles often require resources. A customized resource provision scheme required improve utilization of multi-dimensional In this work, a slice selection-based online offloading (SSOO) algorithm proposed in First, response time energy consumption are reduced processing locally on vehicles. Then,...

A new approach, named bilateral motion data fusion, was proposed for the analysis of movement symmetry, which takes advantage cross-information between both sides body and processes unilateral at same time.This accomplished using canonical correlation joint independent component analysis. It should be noted that human movements include many categories, cannot enumerated one by one. Therefore, gait rhythm fluctuations healthy subjects patients with neurodegenerative diseases were employed as...

In mobile and edge intelligence systems, federated learning (FL) enables local data training model sharing without obtaining actual from users, which are owners. Data processes performed at the user side with only trained gradients passed to an aggregator, i.e., server. The server continually trains updates corresponding models by collecting gradients. updated delivered back users for improved results. Despite advantages of in preserving privacy, process will consume adequate amount energy...

General Purpose Graphics Processing Units (GPGPU) have been used in high performance computing platforms to accelerate the of scientific applications such as simulations. With increased resources required for large-scale network simulation, one GPU device may not enough memory and computation capacities. It is therefore necessary enhance system scalability by introducing multiple devices. also attractive investigate Multi-GPU This paper describes simulation information propagation on...

With the significant advances of AI technology, intelligent robotic systems have achieved remarkable development and profound effects. To enable massive data transmissionin an efficient reliable way, both high performance andhigh reliability should be taken into account in system design. However, conventional communication middleware used majority autonomous systems, is based on socked-based methods, which always lead to latency. Moreover, some sophisticated utilizes shared memory upon ring...

In many large-scale scientific applications, there may be a compute intensive kernel that largely determines the overall performance of application. Sometimes algorithmic variations available and benefit can then gained by choosing optimal at runtime. However, it is sometimes difficult to choose most efficient as algorithms have varying under different execution conditions. This paper shows how construct set models explore analyze bottleneck an Furthermore, based on models, theoretical...

Simulation has become an important method that is widely used in studying the propagation behaviors during process of viral advertisement diffusion. With increased computing and memory resources required for large-scale network processing, General Purpose Graphics Processing Units (GPGPUs) have been high performance platforms to accelerate simulation performance. In this paper, we show optimized strategies diffusion on a Multi-GPU system. Using our proposed strategies, examine spread...

Simulation has become an important method that is widely used in studying the propagation behaviors during process of viral advertisement diffusion. With increased computing and memory resources required for large-scale network processing, General Purpose Graphics Processing Units (GPGPUs) have been high performance platforms to accelerate simulation performance. In this paper, we show optimized strategies diffusion on a Multi-GPU system. Using our proposed strategies, examine spread...

The fusion of the Internet Things (IoT) with Sixth-Generation (6G) technology has significant potential to revolutionize IoT landscape. With ultra-reliable and low-latency communication capabilities 6G, 6G-IoT networks can transmit high-quality diverse data enhance edge learning. Artificial Intelligence-Generated Content (AIGC) harnesses advanced AI algorithms automatically generate various types content. emergence AIGC integrates networks, facilitating real-time provision customized...

Multi-core architectures are widely used to execute large scale scientific applications with a shared memory parallelism. The locking policy of critical sections is protect the data state if multiple threads simultaneously accessing. Factors such as portion section part parallel code, and number concurrency may affect multi-threaded processing can then become bottleneck constraining overall performance. This paper discusses impact on performance, both in task pattern pattern. According...

There has been a remarkable increase in the speed of AI development over past few years. Artificial intelligence and deep learning techniques are blooming expanding all forms to every sector possible. With emerging intelligent autonomous navigation systems, both memory allocation data movement becoming main bottlenecks inter-process communication procedures, especially supporting various types messages between multiple programming languages. To reduce significant cost, we propose novel...

GPUs have been widely adopted to speedup various throughput-originated applications running on HPC platforms, where typically there are a number of tasks sharing maximize GPU utilization. To facilitate sharing, vendors provide tools, allowing multiple processes concurrently use GPUs. For example, Nvidia provides MPS (Multi-Process Service) managing all achieve high throughput by fully exploiting hardware resources. However, such tool leads undesired single point failure for processes,...