ID 原文 译文
2533 如何提高对未知噪声类型的泛化能力是有监督语音增强方法中亟待解决的重要问题，通过对大量不同类型噪声进行建模，深度神经网络成为了解决该问题的有效手段。 How to improve the generalization ability to unknown noise types is an important problem to be solved ur-gently in supervised speech enhancement approaches. By modeling a large number of types of noise, the deep neural network(DNN)becomes an effective way to solve this problem.
2534 为了进一步提高基于深度神经网络的语音增强方法的泛化能力，本文基于生成式对抗网络(Generative Adversarial Networks，GAN)设计了能够由真实噪声数据生成新的噪声类型的 NoiseGAN; In order to further improve the generalization ability of speech en-hancement approaches based on DNN, this paper designs NoiseGAN based on Generative Adversarial Networks (GAN)togenerate new noise types from real noise data.
2535 通过在训练集中增加生成噪声类型，提高训练集噪声类型的多样性，从而达到提高语音增强模型泛化能力的目的。 By adding generated noise to training set, the diversity of noise types in train-ing set is increased, and thereby the generalization ability of speech enhancement model is improved.
2536 不同结构的网络下的语音增强实验结果表明，本文提出的 NoiseGAN 能够生成新的噪声类型，具备提高训练集噪声类型多样性的能力，有效提高了语音增强模型在未知噪声类型下的泛化能力。 The results of speechenhancement experiments under different structures of networks show that the proposed NoiseGAN can generate new noisetypes, increase the diversity of noise types in training set, and effectively improve the generalization ability of speech en-hancement models under unknown noise types.
2537 随着云计算成为重要的信息基础设施，越来越多的应用迁移到云上，云服务的可靠性日益重要，尤其是边缘计算新模式的引入，对云服务可靠性提出了更高的要求。 As Cloud Computing becomes an important information infrastructure, more and more applications are be-ing migrated to the cloud. Therefore, the reliability of cloud services becomes increasingly important. In particular, the intro-duction of new edge computing mode puts forward higher requirements on the reliability of cloud services.
2538 如何通过资源调度保障服务可靠性成为了当前研究的热点。 How to guaranteethe reliability of services through resource scheduling has become a hot topic of current research.
2539 为此，针对云-边协同的应用场景，开展面向服务可靠性的云资源调度方法研究，提出基于马尔科夫预测模型的云资源调度算法，实现节点负载判断、待迁移任务和节点选择、迁移路由的决策，以解决云服务节点失效情况下的任务调度和负载均衡问题， In Cloud-Edge collaborativeapplication scenarios, we research on a service reliability oriented cloud resource scheduling method to support cloud servicereliability. And the cloud resource scheduling algorithm based on markov prediction model is put forward to solve the prob-lem of task scheduling and load balancing in cloud service node failure situation, including the judgment of node load de-gree, the selection of migrated task and nodes, and the decision of migration routing.
2540 实现快速的云服务故障恢复，提高云服务的可靠性。 The goal is to achieve rapid cloud serv-ice recovery and to improve the reliability of cloud services.
2541 实验结果表明，本文所提方法能够有效保证节点失效情况下的服务可靠性。 The experimental results show that the proposed method can ef-fectively guarantee the service reliability.
2542 基于模型的诊断为人工智能领域中一个重要的研究分支，极小碰集即候选诊断的求解过程极大影响最终的诊断效率。 Model-based diagnosis is an important branch of research in the field of artificial intelligence. The efficien-cy for generating all minimal hitting sets, i. e. , candidate diagnoses, considerably affects the final diagnostic process.