Vehicle Edge computing (VEC) has emerged as an efficacious paradigm that supports real-time, computation-intensive vehicular applications. However, due to the highly dynamic nature of computing node topology, existing...
详细信息
As individuals have become overloaded with information, Recommender Systems (RS) were created to provide machine generated recommendations. Significant advancements in RS have been made thanks to Machine Learning meth...
详细信息
Deep Neural networks (DNNs) have become increasingly computationally intensive and have larger parameters, requiring efficient parallelization or distribution using multiple accelerators. Pipeline parallelism has been...
详细信息
ISBN:
(纸本)9798400708893
Deep Neural networks (DNNs) have become increasingly computationally intensive and have larger parameters, requiring efficient parallelization or distribution using multiple accelerators. Pipeline parallelism has been proposed as an effective way to distribute models and improve hardware utilization. However, the problem with pipeline parallelism is the trade-off between speedup and accuracy: synchronous approaches do not provide sufficient speedup, while asynchronous approaches suffer from accuracy degradation due to a different scheme from a single worker. In this paper, we propose AshPipe, a hybrid parallel framework that combines data parallelism and asynchronous pipeline parallelism to achieve efficient speedup for training. The proposed runtime uses the 1F1B schedule and data parallelism, with non-uniform numbers of workers and identical global batch sizes across stages. A Switch parallelism (SP) mechanism is also proposed as an option to mitigate accuracy degradation, which switches over from data parallelism to hybrid parallelism in the course of training. Experimental results show that AshPipe achieves 1.844x the throughput of data parallelism for ViT-H/14 whose parameter size is 632M. With the SP mechanism, AshPipe achieved a 30.2% reduction in training time with comparable accuracy compared to data parallelism when training on the CIFAR100 dataset.
WSN consist of tiny sensors that are distributed over the entire network and have the capability of sensing the data, processing it, and conveying it from one node to another. The purpose of the study is to minimize t...
详细信息
WSN consist of tiny sensors that are distributed over the entire network and have the capability of sensing the data, processing it, and conveying it from one node to another. The purpose of the study is to minimize the power utilization per round and elevate the network lifespan. In the present case, nature-inspired mechanisms are used to minimize the power utilization of the network. In the proposed study, the Butterfly Optimization Algorithm (BOA) is used to choose the optimal quantity of cluster heads from the dense nodes (available nodes). The parameters to be considered for the choice of the cluster head are: the remaining power of the node;distance from the other nodes in the network;distance from the base station;node centrality;and node degree. The particle swarm optimization (PSO) is used to form the cluster head by choosing certain parameters, such as distance from the cluster head and the BS. The path is chosen by means of the Ant Colony Optimization (ACO) Mechanism. The route is optimized by the distance, node degree, and the chosen remaining power. The inclusive performance of the projected protocol is measured in terms of stability period, quantity of active nodes, data acknowledged by the base station, and overall power utilization of the network. The results of the put redirect methodology are correlated with the extant mechanisms such as LEACH, DEEC, DDEEC, and EDEEC (Khan et al. in World Appl Sci J, 2013;Arora and Singh in Soft Comput 23:715-734, 2019;Saini and Sharma in 2010 First internationalconference on parallel, distributed and grid computing (PDGC 2010), 2010;Elbhiri et al. in 2010 5th international symposium on I/V communications and mobile network, 2010) and correlated with the swarm mechanisms such as CRHS, BERA, FUCHAR, ALOC, CPSO, and FLION. This review will help investigators discover the applications in this arena.
The ease with which deep learning can generate fake images has created a pressing need for a robust platform to distinguish between real and fake imagery. However, existing methods in image forensics rely on complex d...
详细信息
The number of Web services on the Internet has been steadily increasing in recent years due to their growing popularity. Under the big data environment, how to effectively manage Web services is of significance for se...
详细信息
The amazing success of deep neural network benefits from the rise of big data. As deep learning models are becoming more scale than ever before, their requirements for memory bandwidth are growing at a tremendous pace...
详细信息
In parallel and distributedcomputing, there are practically two networks: linear networks (also called paths) and rings (also called loops). Many efficient algorithms, such as signal and image processing, were first ...
详细信息
High-performance computing communities are increasingly adopting Neural networks (NN) as surrogate models in their applications to generate scientific insights. Replacing an execution phase in the application with NN ...
详细信息
ISBN:
(纸本)9798400701559
High-performance computing communities are increasingly adopting Neural networks (NN) as surrogate models in their applications to generate scientific insights. Replacing an execution phase in the application with NN models can bring significant performance improvement. However, there is a lack of tools that can help domain scientists automatically apply NN-based surrogate models to HPC applications. We introduce a framework, named Auto-HPCnet, to democratize the usage of NN-based surrogates. Auto-HPCnet is the first end-to-end framework that makes past proposals for the NN-based surrogate model practical and disciplined. Auto-HPCnet introduces a workflow to address unique challenges when applying the approximation, such as feature acquisition and meeting the application-specific constraint on the quality of final computation outcome. We show that Auto-HPCnet can leverage NN for a set of HPC applications and achieve 5.50x speedup on average (up to 16.8x speedup and with data preparation cost included) while meeting the application-specific constraint on the final computation quality.
In this study, we propose BreathPass, a non-invasive authentication system that characterizes the chest/abdomen movement incurred by human breath to enable unlocking smart devices while wearing various types of face c...
详细信息
暂无评论