Internet of Things (IoT) generates large amounts data at the network edge. Machine learning models are often built on these data, to enable detection, classification, and prediction future events. Due bandwidth, storage, especially privacy concerns, it is impossible send all IoT center for centralized model training. To address issues, federated has been proposed let nodes use local train models, which then aggregated synthesize a global model. Most existing work focused designing algorithms...

We are living in a world where massive end devices perform computing everywhere and everyday. However, these constrained by the battery computational resources. With increasing number of intelligent applications (e.g., augmented reality face recognition) that require much more power, they shift to computation offloading cloud, known as mobile cloud (MCC). Unfortunately, is usually far away from devices, leading high latency well bad quality experience (QoE) for latency-sensitive...

Neural machine translation (NMT) aims at solving (MT) problems using neural networks and has exhibited promising results in recent years. However, most of the existing NMT models are shallow there is still a performance gap between single model best conventional MT system. In this work, we introduce new type linear connections, named fast-forward based on deep Long Short-Term Memory (LSTM) networks, an interleaved bi-directional architecture for stacking LSTM layers. Fast-forward connections...

The emergence of edge computing has witnessed a fast-growing volume data on devices belonging to different stakeholders which, however, cannot be shared among them due the lack trust. By exploiting blockchain's non-repudiation and non-tampering properties that enable trust, we develop blockchain-based big sharing framework support various applications across resource-limited edges. In particular, devise number novel resource-efficient techniques for framework: (1) PoC...

Federated learning is promising in enabling large-scale machine by massive clients without exposing their raw data. It can not only enable the to preserve privacy information, but also achieve high performance. Existing works of federated mainly focus on improving performance terms model accuracy and task completion time. However, practice, are reluctant participate process receiving compensation. Therefore, how effectively motivate actively reliably paramount. As compared current incentive...

Edge computing is a new paradigm to provide strong capability at the edge of pervasive radio access networks close users. A critical research challenge design an efficient offloading strategy decide which tasks can be offloaded servers with limited resources. Although many efforts attempt address this challenge, they need centralized control, not practical because users are rational individuals interests maximize their benefits. In article, we study decentralized algorithm for computation...

Due to digitalization, small and medium-sized enterprises (SMEs) have significantly enhanced their efficiency productivity in the past few years. The process automate SME transaction execution is getting highly multifaceted as number of stakeholders SMEs connecting, accessing, exchanging, adding, changing transactional executions. balanced lifecycle requires partnership exchanges, financial management, manufacturing, stabilities, along with privacy security. Interoperability platform issue...

Multicast is an important mechanism in modern wireless networks and has attracted significant efforts to improve its performance with different metrics including throughput, delay, energy efficiency, etc. Traditionally, ideal loss-free channel model widely used facilitate routing protocol design. However, the quality of links affected or even jeopardized resulting transmission failures by many factors like collisions, fading noise environment. In this paper, we propose a reliable multicast...

The explosive growth of demands on big data processing imposes a heavy burden computation, storage, and communication in centers, which hence incurs considerable operational expenditure to center providers.Therefore, cost minimization has become an emergent issue for the upcoming era.Different from conventional cloud services, one main features services is tight coupling between computation as tasks can be conducted only when corresponding are available.As result, three factors, i.e., task...

Federated learning is promising in enabling large-scale machine by massive mobile devices without exposing the raw data of users with strong privacy concerns. Existing work federated struggles for accelerating process, but ignores energy efficiency that critical resource-constrained devices. In this paper, we propose to improve lowering CPU-cycle frequency who are faster training group. Since all synchronized iterations, speed preserved as long they complete before slowest device each...

Sharding can significantly improve the blockchain scalability, by dividing nodes into small groups called shards that handle transactions in parallel. However, all existing sharding systems adopt complete sharding, i.e., are isolated. It raises additional overhead to guarantee atomicity and consistency of cross-shard seriously degrades performance. In this paper, we present Pyramid, first layered system, which some store full records multiple thus be processed validated these internally....

In the age of big data, companies tend to deploy their services in data centers rather than own servers. The demands analytics grow significantly, which leads an extremely high electricity consumption at centers. this paper, we investigate cost minimization problem on geo-distributed connected renewable energy sources with unpredictable capacity. To solve problem, propose a Reinforcement Learning (RL) based job scheduling algorithm by combining RL neural network (NN). Moreover, two...

Wearable computing becomes an emerging paradigm for various recently developed wearable devices, such as Google Glass and the Samsung Galaxy Smartwatch, which have significantly changed our daily life with new functions. To magnify applications on devices limited computational capability, storage, battery capacity, in this paper, we propose a novel three-layer architecture consisting of mobile remote cloud code offloading. In particular, offload portion computation tasks from to local or...

Knowledge selection plays an important role in knowledge-grounded dialogue, which is a challenging task to generate more informative responses by leveraging external knowledge. Recently, latent variable models have been proposed deal with the diversity of knowledge using both prior and posterior distributions over achieve promising performance. However, these suffer from huge gap between selection. Firstly, module may not learn select properly because lacking necessary information. Secondly,...

Federated learning has attracted much research attention due to its privacy protection in distributed machine learning. However, existing work of federated mainly focuses on Convolutional Neural Network (CNN), which cannot efficiently handle graph data that are popular many applications. Graph (GCN) been proposed as one the most promising techniques for learning, but setting seldom explored. In this article, we propose FedGraph among multiple computing clients, each holds a subgraph....

In this paper, a food anti-counterfeiting traceability system based on blockchain and the Internet of Things is proposed in response to problems data centre-based storage, easy tampering silos traditional technology. The makes use decentralized storage untamperable characteristics technology store process production, sale transportation, so as ensure uniqueness food. At same time, through things authenticity reliability source blockchain. experimental results show that has higher security,...

As a promising solution to blockchain scalability, sharding divides nodes into small groups called shards, splitting the workload. Existing works for sharding, however, are limited by cross-shard transactions, since they need split each transaction multiple sub-transactions, of which costs consensus round commit. In this paper, we introduce PYRAMID, novel system based on idea layered sharding. with better hardware allowed participate in shards and store blockchains these thus can validate...

Software-Defined Network (SDN) is a promising network paradigm that separates the control plane and data in network. It has shown great advantages simplifying management such new functions can be easily supported without physical access to switches. However, Ternary Content Addressable Memory (TCAM), as critical hardware storing rules for high-speed packet processing SDN-enabled devices, supplied each device with very limited quantity because it expensive energy-consuming. To efficiently use...

Big data analytics has attracted close attention from both industry and academic because of its great benefits in cost reduction better decision making. As the fast growth various global services, there is an increasing need for big across multiple centers (DCs) located different countries or regions. It asks support a cross-DC processing platform optimized geo-distributed computing environment. Although some recent efforts have been made analytics, they cannot guarantee predictable job...