ID 原文 译文
56188 首先基于ratio association和ratio cut,提出两种更适合社区隐藏的指标; Firstly, We model community deception as network growth and introducetwo measures based on ratio association and ratio cut.
56189 然后基于强化学习框架,定义动作空间为不同的网络增长模型,将两个指标在l个阶段的一致性策略值作为网络的状态表示,并将指标值作为奖赏值; Then, based on the reinforcement learning framework,we define the action space as the different network growth models, state as the performance vector of the agent’slatest l strategies and reward as the values of two measures.
56190 最后采用两种策略来对指标进行优化,即给每个Q函数赋予权重的标量化多目标Q-learning算法以及基于Pareto最优算法的多目标Q-learning算法. Finally, we adopt two strategies to optimize the twoobjectives, i. e. , scalarized multi-objective Q-learning and pareto-optimized multi-objective Q-learning.
56191 在真实数据集上的大量实验表明,相比于现有最新的社区隐藏算法,本文所提算法展现出更好的有效性. Extensiveexperiments on real-world datasets demonstrate that the proposed approach can achieve better performance thanlatest community deception methods.
56192 针对一类离散时间非线性动态系统,把常规广义预测控制与数据驱动控制相结合、反馈控制与前馈补偿相结合、信号补偿与调节器原理相结合,提出了一种数据与虚拟未建模动态及其增量补偿的非线性广义预测控制方法. For a class of discrete-time nonlinear dynamic systems, a nonlinear generalized predictive controlmethod is proposed, which combines the conventional generalized predictive control with data-driven control, thefeedback control with feedforward compensation, and the signal compensation with the regulator principle.
56193 推导了闭环系统的输入输出动态方程,并分析了所提控制算法的性能. Theinput-output dynamic equations of the closed-loop system are derived, and the performance of the proposed controlalgorithm is analyzed.
56194 对比实验的研究体现出本文所提控制算法的优越性和有效性. The advantages and effectiveness of the proposed control algorithm are demonstratedthrough comparative experiments.
56195 群智感知系统中针对数据流的实时发布和深度学习在极大方便人们日常生活的同时,也严重威胁了参与用户的隐私信息. The real-time publishing and deep exploitation of data streams in crowdsensing systems have signif?icantly facilitated people’s daily lives.
56196 现有隐私保护机制在处理动态性强、时空相关性复杂的数据流时,大都难以实现数据自适应性,从而导致较低的数据效用性. However, it also seriously compromises the private information of partic?ipating users. The existing approaches are non-adaptive to dynamic changes of streams, thus are vulnerable tolow data utility.
56197 因此,基于ω-事件级差分隐私,本文提出了一种数据自适应的多维数据流隐私保护实时发布机制AdaPub. To address such concerns, in this paper, we present AdaPub, a data-adaptive mechanism forinfinite multi-dimensional stream real-time publishing under ω-event differential privacy.