This paper focuses on an important research problem of Big Data classification in intrusion detection system. Deep Belief Networks is introduced to the field detection, and model based proposed apply recognition domain. The deep hierarchical a neural network classifier combination multilayer unsupervised learning networks, which called as Restricted Boltzmann Machine, supervised network, Back-propagation network. experimental results KDD CUP 1999 dataset demonstrate that performance better...

In daily life, people carry smartphones every where. The sensors included in can tell us much information. Activity recognition by smartphone be used for healthcare and sports management. People different positions, such as the pocket of trousers, hands or bags. We use accelerometer embedded to classify five activities, staying still, walking, running, going upstairs downstairs. This work analysis behavior data from accelerometer, extract various features, choose highly correlated construct...

Web browsing is an activity that billions of mobile users perform on a daily basis. Battery life primary concern to many who often find their phone has died at most inconvenient times. The heterogeneous multi-core architecture solution for energy-efficient processing. However, the current web browsers rely operating system exploit underlying hardware, which no knowledge individual contents and leads poor energy efficiency. This paper describes automatic approach render workloads performance...

Interference management and energy are two important issues in ultra-dense heterogeneous cellular networks (HetNets). However, load balancing investigated for system capacity on interference coordination is not efficient saving HetNets. Meanwhile, the maximum efficiency configuration leads to serious unfair problem users associated with larger power node scenario. Thus, this paper, we investigate consumption jointly together HetNets, formulate max-min energy-efficient enhanced intercell...

In preclinical researches, bioluminescence tomography (BLT) has widely been used for tumor imaging and monitoring, imaged-guided therapy, so forth. For these biological applications, both spatial location morphology analysis are the leading problems needed to be taken into account. However, most existing BLT reconstruction methods were proposed some specific applications with a focus on sparse representation or recovery, respectively. How design versatile algorithm that can simultaneously...

The densification and expansion of heterogeneous cellular networks (HetNets) pose new challenges on interference management reduction energy consumption. 3GPP has proposed enhanced intercell coordination (eICIC) by making a macrocell silent in almost blank subframes (ABSs) to mitigate for low power base stations (BSs) HetNets. However, efficiency (EE) is very crucial the deployment large number nodes as they consume lot energy. In this work, we develop novel EE-eICIC algorithm determine...

Dynamic TDD is one important feature of 3GPP Release 12. In parallel with the standardization, a testbed has been developed to verify feasibility dynamic and evaluate performance in different scenarios. Both dynamic-TDD capable UE eNode B have developed. Fast configuration using explicit physical layer signaling successfully implemented over air measurements conducted indoor scenario cases. The traffic variation based adaptation switch can be done within 10 ms. field measurement results show...

Interference management and power transfer can provide a significant improvement over the 5th generation mobile networks (5G) dense Internet of Things (IoT) heterogeneous (HetNets). In this paper, we present novel approach to simultaneously manage inferences at downlink (DL) uplink (UL), identify opportunities for additional UL transmissions integrated with existing protocols infrastructures enhanced inter-cell interference coordination (eICIC) protocol in IoT HetNets, while considering...

A trusted execution environment (TEE) is a system-on-chip and CPU system with wide security solution available on today's Arm application (APP) processors, which dominate the smartphone market.Generally, mobile APPs create (TA) in TEE to process sensitive information, such as payment or message encryption, transparent running rich environments (REEs).In detail, REE interact eventually send back results APP through interface provided by TA.Such an operation definitely increases overhead of...

It has been shown in the existing literature that data symbols can assist massive MIMO transmission by relieving issue of pilot contamination. In order to further evaluate performance gain from system point view, a stochastic geometry based framework is established this paper analyze distribution signal-to-interference ratio cellular network with data-assisted uplink detection scheme. The closed-from expressions asymptotic bounds are thereby derived. numerical simulation analytical fit real...

With the large-scale promotion of cloud computing, intrusion detection is a necessary way to guaranteen security. However, because lack adaptive model, accuracy still challenge issue in detection. In our works, based on attribute significant expressing cooperative deep belief network (CDBN) was proposed for specific attack Firstly, multi-view division method extract features attack. secondly, an coding mechanism encoding described denoise and compress features.finally, cumulative prospects,...

Intrusion detection has always been a hot and difficult topic in the field of computer security. It is to use traditional intrusion methods effectively detect unknown types. To solve this difficulty, paper, zero-shot method based on regression model proposed identify types order provide guarantee for The includes firstly taking data normal state known type as training set. If features are non-numeric, one-hot code used convert non-numeric into numerical features. In addition, overcome...

In this paper, we propose a new parallel statistical analysis method for large analog circuits using determinant decision diagram (DDD) based graph technique on GPU platforms. DDD-based symbolic enables exact of vary circuits. But show that is very amenable massively threaded computing We design novel data structures to represent the DDD graphs in GPUs enable fast memory access massive threads numerical values graphs. The inspired by inherent parallelism and simple independence evaluation...

Video action recognition is an important research content in the field of computer vision. However, single feature motion information under-represented and cannot completely describe information. In this paper, method based on parallel multi-granularity refinement network was propose to improve accuracy. This relaxes requirement restriction by describing a video with multiple class labels shared features different group. Three granularity obtained three label groups integrate them obtain...

The limited capacity battery of smartphones always incurs low-battery anxiety for mobile users, especially the high energy consumption applications, such as games and videos. They need to render complex content in real-time transfer heavy data from server, leading consumption. As a case study, we investigate Mobile Augmented Reality (MAR), promising technique that has been applied many applications. MAR can provide an immersive user experience by rendering real-world scenarios with...

Summary Deep learning has made great achievements in the field of speech recognition. With popularization embedded devices such as intelligent speaker and demand for dialect interaction scenes, it poses challenges to far‐field recognition language In order solve recognition, we propose a deep neural network model with multitask learning. First, audio is passed through end‐to‐end noise reduction improve effect Then define main task area auxiliary task, using method accuracy classification....