ID	原文	译文
3343	同时，通过在数据集上的批次训练及增加扰动范数约束，提升了对抗补丁的泛化性和多场景可用性。	Meanwhile, throughbatch-training on the dataset and increasing perturbation norm constraint, the generalization and the multi-scene usabilityof the adversarial patch were improved.
3344	在 ILSVRC2012 数据集上的实验结果表明，与现有方法相比，所提方法能够在保持模型分类精度的同时，更简单有效地攻击Grad-CAM 解释结果。	Experimental results on the ILSVRC2012 dataset show that compared with the existing methods, the proposed method can attack the interpretation results of the Grad-CAM more simply and effectively while maintaining the classification accuracy.
3345	为了解决已有图注意力机制在进行链接预测相关任务时，容易造成注意力分配向某些出现频率高的关系倾斜的问题，提出了一种基于层级注意力机制的链接预测模型。	In order to solve the problem that the existing graph attention mechanism tends to cause attention distribution to certain relations with high frequency when performing link prediction related tasks, a new link prediction model basedon hierarchical attention mechanism was proposed.
3346	在链接预测任务中，通过设计分层注意力机制，根据预测任务中的关系对知识图谱中与给定实体相连的不同类型的关系给予不同的注意力。	In the link prediction task, a hierarchical attention mechanism wasdesigned to give different attention to the relationships of different relationship types connected to a given entity in theknowledge graph according to the relationship in the prediction task.
3347	在关注多跳邻居实体特征的同时，更关注关系特征以找到符合目标关系的关系类型。	While the characteristics of multi-hop neighbor enti-ties were pay attention to, the relationship characteristics was pay more attention to find the relationship type that matchesthe target relationship.
3348	在多个基准数据集上与主流模型进行对比实验，实验结果表明，所提模型性能优于主流模型，并具有较好的稳健性。	Through comparison experiments with the mainstream models on multiple benchmark data sets,the results show that the performance of the model is better than the mainstream models and has good robustness.
3349	根据 Wi-Fi 和 LTE 的频谱划分决策分别取决于全局信道状态信息（CSI）和局部 CSI，提出了双时间尺度频谱划分算法，在较大时间尺度上根据全局 CSI 划分 Wi-Fi 频谱，在较小时间尺度上根据局部 CSI 划分 LTE 频谱。	According to the spectrum partitioning decisions of Wi-Fi and LTE respectively depended on global channelstate information (CSI) and local CSI, an online two-timescale iteration algorithm was developed. The Wi-Fi spectrumwas partitioned at the large timescale according to the global CSI whereas the LTE spectrum was partitioned according to the fast-changing local CSI.
3350	设计了自适应反馈补偿机制以提高双时间尺度算法的跟踪性能。	Then, the adaptive compensation mechanism was designed for improving the tracking per-formance of the two-timescale algorithm.
3351	此外，得到双时间尺度算法能够无差错地跟踪上均衡点的充分条件。	Moreover, the sufficient condition was derived for the two-timescale algorithmtracking the moving equilibrium point without errors.
3352	仿真实验表明，双时间尺度算法能够以非常低的开销实现较好系统性能。	Finally, the simulation results show that the proposed two-time scale algorithm achieves excellent system performance at very low overhead.