Recent advances in computing have allowed for the possibility to collect large amounts of data on personal activities and private living spaces. To address the privacy concerns of users in this environment, we propose a novel framework called PR-GAN that offers privacy-preserving mechanism using generative adversarial networks. Given a target application, PR-GAN automatically modifies the data to hide sensitive attributes -- which may be hidden and can be inferred by machine learning algorithms -- while preserving the data utility in the target application. Unlike prior works, the public's possible knowledge of the correlation between the target application and sensitive attributes is built into our modeling. We formulate our problem as an optimization problem, show that an optimal solution exists and use generative adversarial networks (GAN) to create perturbations. We further show that our method provides privacy guarantees under the Pufferfish framework, an elegant generalization of the differential privacy that allows for the modeling of prior knowledge on data and correlations. Through experiments, we show that our method outperforms conventional methods in effectively hiding the sensitive attributes while guaranteeing high performance in the target application, for both property inference and training purposes. Finally, we demonstrate through further experiments that once our model learns a privacy-preserving task, such as hiding subjects' identity, on a group of individuals, it can perform the same task on a separate group with minimal performance drops.
As one of the significant driving forces of the future broadband heterogeneous networks (5G), the Next Generation WLAN (NGW) has been initiated following the 802.11 HEW (High-Efficiency WLAN) study group with a focus on improving spectrum efficiency and area throughput. However, Low-Rate-Links (LRLs) severely degrade the overall wireless network performance, particularly under the dense environments in the NGW. To improve the data rate of wireless links, cooperation communication is proposed to divide a low-rate-link into multiple successive wireless links with much higher data rate over the same frequency channel, which is denoted as `Serial-Coop'/`SC' strategy. However, the existing `Serial-Coop' strategy suffers from the disadvantage of large cumulative delay because the packets are relayed within two successive timeslots. In this paper, a distributed Multi-channel MAC protocol with Parallel Cooperation is proposed for the NGW, called PC-MMAC, to initiate concurrent cooperation communications over multiple frequency channels in a distributed fashion. Then, an approximate analysis on the upper bound of saturation throughput gain is derived. Further, an enhanced mechanism is proposed to strengthen the robustness of PC-MMAC, by dynamically adjusting the cooperation strategy based on the number of available data channels. Extensive simulation results show that the saturation throughput of PC-MMAC outperforms `Serial-Coop' strategy and `Non-Coop' scheme 1 by 25% and 33% respectively.
To study how to improve the efficiency of cruise phase of TV-guided air-to-surface missile,a calculation model was proposed,which consisted of penetration probability and target acquisition probability.Then research was made on the effect that altitude and velocity of cruise phase had on the penetration probability and target acquisition probability of the missile.Finally,simulation was made for the efficiency of cruise phase of the missile by using the method of Monte Carlo.The result showed that when altitude of cruise phase was between 160 m and 360 m the efficiency was higher and the faster the missile was,the higher the efficiency was.
The BIST circuit based on March algorithm can reach to very high fault coverage,but it exposes the shortcoming of occupying too large area when testing SRAM of small scale .In this article,based on the working characteristic of the four pieces of 640×18 bit SRAM in timing controller LTTC1 of big panel,a new test method was brought forward. The new method imposed test pattern on the SRAM according to continuous and incremental address,avoiding the unnecessary test led to by March C. Meanwhile,it made the MBIST circuit superior, reducing the number of transistors and area overhead led to by routing. The results showed that it can produce the test effect that March C+ can do.
IEEE1588 is a new kind of time synchronization protocol that is suitable for distributed instruments and can achieve higher precision.The basic theory of the IEEE1588 time synchronization and the features of the IEEE1588 that are different from other protocols are demonstrated in the paper.At the same time,the hardware design of IEEE1588 and its application in the system of LXI instruments are put forward.Finally,the expectation about the development of the IEEE1588 in the future is discussed.
An adaptive beam forming algorithm with antenna array model based on quantum-behaved particle swarm optimization (QPSO) is proposed for improving existent problems of limited beam modulation, unsatisfactory side lobe suppression and poor adaptability. By adjusting beam direction with matrix-weighting, optimizing initial phase and radius of the antenna array, and by tracking user intelligently with amplify and forward (AF) technology accordingly, better suppressing interference and economizing energy can be realized to effectively increase the received power and make use of antenna array for improving practical application. The simulation results of MATLAB show that the gain rising about 10dB and interference declining about 3.75dB at the direction of main lobe can be achieved.
In order to improve the real-time processing speed of the adaptive filter and the algorithm stability,the latest DSP Builder software is adopted to realize uncorrelated normalized LMS(DNLMS) algorithm based on FPGA.This method is proved to reach good effect by simulation in Matlab/Simulink.The uncorrelated principle and the process of normalization is substituted,the convergence speed could be quickened and the misadjustment could be smaller.If this adaptive is added from 8 stages to 512 stages or 1 024 stages,noise elimination in speech processing or other-adaptive filter is well realized.
Based on the current standardized IEEE 802.11 distributed coordination function (DCF) protocol, this paper proposes a new efficient collision resolution mechanism, called GDCF (gentle DCF). Our main motivation is based on the observation that 802.11 DCF decreases the contention window to the initial value after each success transmission, which essentially assumes that each successful transmission is an indication that the system is under low traffic loading. GDCF takes a more conservative measure by halving the contention window size after c consecutive successful transmissions. This "gentle" decrease lowers the collision probability, especially when the competing node number is large. We compute the optimal value for c, and the numerical results from both analysis and simulation demonstrate that GDCF significantly improve the performance of 802.11 DCF including throughput, fairness, and energy efficiency. In addition, GDCF is flexible for supporting priority access by selecting different values of c for different traffic types; it is fully compatible with the original 802.11 DCF, and simple to implement, as it does not requires changes in control message structure and access procedures in DCF.
How can animals behave effectively in conditions involving different motivational contexts? Here, we propose how reinforcement learning neural networks can learn optimal behavior for dynamically changing motivational salience vectors. First, we show that Q-learning neural networks with motivation can navigate in environment with dynamic rewards. Second, we show that such networks can learn complex behaviors simultaneously directed towards several goals distributed in an environment. Finally, we show that in Pavlovian conditioning task, the responses of the neurons in our model resemble the firing patterns of neurons in the ventral pallidum (VP), a basal ganglia structure involved in motivated behaviors. We show that, similarly to real neurons, recurrent networks with motivation are composed of two oppositely-tuned classes of neurons, responding to positive and negative rewards. Our model generates predictions for the VP connectivity. We conclude that networks with motivation can rapidly adapt their behavior to varying conditions without changes in synaptic strength when expected reward is modulated by motivation. Such networks may also provide a mechanism for how hierarchical reinforcement learning is implemented in the brain.
With the rapid development of mobile internet services, the intensive deployment of wireless local area network (WLAN) is inevitable. Traditional WLAN uses carrier sense/collision avoidance (CSMA/CA) mechanism to avoid interference between links as much as possible. In the multi access points (APs) scenario, the traditional CSMA/CA may lead to the flow in the middle (FIM) problem, resulting in a sharp reduction in the throughput of the intermediate nodes and affecting the fairness of the whole network. Existing researches have shown that the FIM problem is more serious in the high density WLAN network. The existing researches on the FIM problem mainly focus on the dynamic optimization of a single parameter, the fairness and performance of the network cannot be well guaranteed. Therefore, this paper proposes a FIM oriented down link (DL) multi-parameter joint dynamic control scheme. AP as a centralized controller, regularly obtains the transmission status of the whole network, reduces the transmission opportunities of the strong AP through adaptive dynamic power and energy detection threshold control (A-DPEC) algorithm, and improves the transmission opportunities of the starvation AP, so as to achieve the performance and fair balance of the whole network. The simulation results show that the proposed scheme outperforms the comparing schemes.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)