Image-text matching is an important problem at the intersection of computer vision and natural language processing. It aims to establish the semantic link between image and text to achieve high-quality semantic alignm...
详细信息
Recently, AI and deep neural networks have found extensive applications in mobile devices, drones, carts, and more. To meet the demands of processing large-scale data and providing DNN inference services with minimal ...
详细信息
ISBN:
(纸本)9798400716713
Recently, AI and deep neural networks have found extensive applications in mobile devices, drones, carts, and more. To meet the demands of processing large-scale data and providing DNN inference services with minimal latency, there is a need. However, IoT devices, with their limited computing capabilities, are not well-suited for AI inference. Moreover, considering the diverse requirements of different services, it is necessary to provide inference services that address these variations. To address these challenges, many previous studies have explored collaborative approaches between edge servers and cloud servers by partitioning DNN models. However, these methods face difficulties in finding optimal partitioning points for splitting DNN models and are heavily influenced by network bandwidth since intermediate computation results need to be transmitted to other devices. In this paper, we propose the Adaptive block-based DNN network inference framework. This involves breaking down a large DNN model into block-level networks, training them using knowledge distillation techniques to enable inference only through each block network. Subsequently, dynamic block-level inference calculations are offloaded based on the computing capabilities of edge clusters, providing inference results. Even when using multiple devices, our method is not affected by network bandwidth since only input images need to be transmitted. Experimental results demonstrate that our approach consistently reduces inference latency as the number of devices increases. Additionally, by controlling the trade-off between latency and accuracy, we can provide inference services tailored to various latency requirements.
Deep neural networks (DNNs) have gained popularity in various fields, including computer vision and speech recognition. However, the growing size of data and complexity of models have made training DNNs increasingly t...
详细信息
With the increasing proliferation of Internet of Things (IoT) and industrial big data technologies, while IoT scenarios offer intelligent services, cyberspace faces various types of threats and attacks. Despite Graph ...
详细信息
Influence maximization (IM) is a fundamental operation among graph problems that involve simulating a stochastic diffusion process on real-world networks. Given a graph G(V, E), the objective is to identify a small se...
详细信息
ISBN:
(纸本)9781665494236
Influence maximization (IM) is a fundamental operation among graph problems that involve simulating a stochastic diffusion process on real-world networks. Given a graph G(V, E), the objective is to identify a small set of key influential "seeds"i.e., a fixed-size set of k nodes, which when influenced is likely to lead to the maximum number of nodes in the network getting influenced. The problem has numerous applications including (but not limited to) viral marketing in social networks, epidemic control in contact networks, and in finding influential proteins in molecular networks. Despite its importance, application of influence maximization at scale continues to pose significant challenges. While the problem is NP-hard, efficient approximation algorithms that use greedy hill climbing are used in practice. However those algorithms consume hours of multithreaded execution time even on modest-sized inputs with hundreds of thousands of nodes. In this paper, we present IMpart, a partitioning-based approach to accelerate greedy hill climbing based IM approaches on both shared and distributed memory computers. In particular, we present two parallel algorithmsone that uses graph partitioning (IMpart-metis) and another that uses community-aware partitioning (IMpart-gratis)with provable guarantees on the quality of approximation. Experimental results show that our approaches are able to deliver two to three orders of magnitude speedup over a state-of-the-art multithreaded hill climbing implementation with negligible loss in quality. For instance, on one of the modest-sized inputs (Slashdot: 73K nodes;905K edges), our partitioning-based shared memory implementation yields 4610x speedup, reducing the runtime from 9h 36m to 7 seconds on 128 threads. Furthermore, our distributed memory implementation enhances problem size reach to graph inputs with x10(6) nodes and x10(8) edges and enables sub-minute computation of IM solutions.
The solution of over-determined equations plays a very important role in fields such as data fitting, signal processing, and machine learning. It is of great significance in predicting natural phenomena, optimizing en...
详细信息
The power line is a crucial component of the power system. However, long-term exposure to the natural environment can lead to various defects, such as burrs, corrosion, and corona. These defects seriously threaten the...
详细信息
Quick genomic data transfer acts as an essential tool that public health teams need during emergencies like the COVID-19 outbreak. Network coding stands as a technology evaluated for improving large-scale genomic data...
详细信息
The increasing volume of Internet traffic is pushing the Internet Service Providers to deploy distributed caching services at the network edge, close to the end users, in order to speed up the data retrieval and reduc...
详细信息
ISBN:
(纸本)9781538674628
The increasing volume of Internet traffic is pushing the Internet Service Providers to deploy distributed caching services at the network edge, close to the end users, in order to speed up the data retrieval and reduce the bandwidth demands. In parallel, centralized paradigms like Software Defined Networking (SDN) are considered to optimize network management while supporting a variety of network applications like routing, load balancing and caching. In this paper, we extend the SDN control plane to support a novel content caching strategy. We consider a softwarised edge network domain where SDN nodes, augmented with storage capabilities, cache incoming data with the twofold target of limiting the retrieval delay and the inter-domain traffic. The caching decision is taken in a centralized mode by the SDN Controller, according to a newly defined content-driven closeness centrality metric, which identifies the importance of the SDN nodes as cachers based on their proximity to the majority of the clients requesting the most popular contents. Simulation results show the superiority of the solution in terms of higher cache hits and reduced latency, when compared against benchmark caching strategies.
With the continuous development of technologies like cloud computing, the network edge is gradually fading away. Zero Trust, as a novel concept in network security, provides new insights into access control. However, ...
详细信息
暂无评论