The use of peer-to-peer (P2P) applications is growing dramatically, particularly for sharing large video/audio files and software. In this paper, we analyze P2P traffic by measuring flow-level information collected at multiple border routers across a ISP network, report our investigation three popular systems-FastTrack, Gnutella, Direct-Connect. We characterize the observed single its impact on underlying network. observe very skewed distribution in network different levels spatial...

Graph-based semisupervised learning (GSSL) provides a promising paradigm for modeling the manifold structures that may exist in massive data sources high-dimensional spaces. It has been shown effective propagating limited amount of initial labels to large unlabeled data, matching needs many emerging applications such as image annotation and information retrieval. In this paper, we provide reviews several classical GSSL methods few handling challenging issues often encountered web-scale...

Face recognition has achieved significant progress with the growing scale of collected datasets, which empowers us to train strong convolutional neural networks (CNNs). While a variety CNN architectures and loss functions have been devised recently, we still limited understanding how models label noise inherent in existing face datasets. To address this issue, paper develops novel co-mining strategy effectively on datasets noisy labels. Specifically, simultaneously use values as cue detect...

Due to the high-energy efficiency and scalability, clustering routing algorithm has been widely used in wireless sensor networks (WSNs). In order gather information more efficiently, each node transmits data its Cluster Head (CH) which it belongs, by multi-hop communication. However, communication cluster brings problem of excessive energy consumption relay nodes are closer CH. These nodes' will be consumed quickly than farther nodes, negative influence on load balance for whole networks....

Speech production is a hierarchical mechanism involving the synchronization of brain and oral articulators, where intention linguistic concepts transformed into meaningful sounds. Individuals with locked-in syndrome (fully paralyzed but aware) lose their motor ability completely including articulation even eyeball movement. The neural pathway may be only option to resume level communication for them. Current brain-computer interfaces (BCIs) use patients' visual attentional correlates build...

In recent years, with the development of intelligent vehicles and wireless sensor network technology, research on road safety has attracted much attention in vehicular ad-hoc networks (VANETs). By sensing events road, can broadcast information to inform others traffic jams or accidents. However, mobile vehicle a large transmission delay, which makes real-time content impossible. this paper, new certificateless aggregate signcryption scheme (CLASC) is proposed by using fog computing framework...

Over the last decade, machine learning (ML) and deep (DL) algorithms have significantly evolved been employed in diverse applications, such as computer vision, natural language processing, automated speech recognition, etc. Real-time safety-critical embedded Internet of Things (IoT) systems, autonomous driving UAVs, drones, security robots, etc., heavily rely on ML/DL-based technologies, accelerated with improvement hardware technologies. The cost a deadline (required time constraint) missed...

Scientific applications are adapting HDFS/MapReduce to perform large scale data analytics. One of the major challenges is that an overabundance small files common in these applications, and HDFS manages all its through a single server, Namenode. It anticipated can significantly impact performance In this work we propose mechanism store efficiently improve space utilization for metadata. Our scheme based on assumption each client assigned quota file system, both number files. our approach,...

With the technology of data storage and shared services in cloud, users can easily share or revoke from their groups. To ensure accuracy data, key is used for group to calculate authenticators audit data's integrity remotely. However, high-speed mobile terminals have posed challenges management Vehicle Ad-hoc networks (VANETs). Besides, maintain computation communication overheads must be increased revocation process when a vehicle user removed group. For such, new auditing scheme proposed...

Visual Question Answering (VQA) is a multimodal task involving Computer Vision (CV) and Natural Language Processing (NLP), the goal to establish high-efficiency VQA model. Learning fine-grained simultaneous understanding of both visual content images textual questions heart VQA. In this paper, novel Multimodal Encoder-Decoder Attention Networks (MEDAN) proposed. The MEDAN consists (MEDA) layers cascaded in depth, can capture rich reasonable question features image by associating keywords...

Recent research works have been presented on conserving energy for multi-disk systems either at a single disk drive level or storage system and thereby having certain limitations. This paper studies several new redundancy-based, power-aware, I/O request scheduling cache management policies the RAID controller to build energy-efficient systems, by exploiting redundant information destage issues of array two popular levels, 1 5. For 1, we develop Windowed Round Robin (WRR) policy; 5, introduce...

Recently, high-energy consumption has become a serious concern for both storage servers and data centers. Recent research studies have utilized the short transition times of multispeed disks to decrease energy consumption. Manufacturing challenges costs so far prevented commercial deployment disks. In this paper, we propose an saving policy, eRAID (energy-efficient RAID), conventional disk-based mirrored parity redundant disk array architectures. saves by spinning down partial or entire...

An efficient and distributed scheme for file mapping or lookup is critical in decentralizing metadata management within a group of servers. This paper presents novel technique called Hierarchical Bloom Filter Arrays (HBA) to map filenames the servers holding their metadata. Two levels probabilistic arrays, namely, filter arrays with different accuracies, are used on each server. One array, lower accuracy representing distribution entire metadata, trades significantly reduced memory overhead,...

This paper presents Manifold, an open-source parallel simulation framework for multicore architectures. It consists of a kernel, set microarchitecture components, and integrated library power, thermal, reliability, energy models. Using the components as building blocks, users can assemble architecture models perform serial or simulations to study architectural and/or physical characteristics Users also create new Manifold port existing Importantly, Manifold's component-based design provides...

To quantify the articulatory distinctiveness of 8 major English vowels and 11 consonants based on tongue lip movement time series data using a data-driven approach.Tongue movements from 10 healthy talkers were collected. First, classification accuracies obtained 2 complementary approaches: (a) Procrustes analysis (b) support vector machine. distance was then used to measure among consonants. Finally, (distinctiveness) matrices different vowel pairs consonant derive spaces multidimensional...

Abstract Salt mist is a significant environmental factor that causes corrosion of marine and coastal engineering facilities electronic equipment failures. This paper presents numerical analysis the properties, formation, migration, deposition processes salt aerosols according to previous research. We find distributed as exponential distribution in sea vertical direction, number them directly proportional wind speed effective wave height, critical velocity sedimentation square particle...

Energy efficiency becomes increasingly important in today's high-performance storage systems. It can be challenging to save energy and improve performance at the same time conventional (i.e. single-rotation-rate) disk-based Most existing solutions compromise for conservation. In this paper, we propose a redundancy-based, two-level I/O cache architecture called RIMAC address problem. The idea of is enable data on standby disk recovered by accessing or currently active/idle disks. At both...