ID	原文	译文
24555	近年来物联网在许多军事和民用领域（灾后恢复、环境监控和军事对抗等）展现出蓬勃的应用前景，而在实际应用中，为了维护终端数据的新鲜度，必须以尽可能低的时延来完成数据聚合调度，从而为用户提供及时准确的数据服务。	Recent years have witnessed booming application of Internet of Things to many military and civilian fields (disaster recovery, environmental monitoring, military confrontation, and so on). In practical application, in order to maintain the freshness of terminal data, data aggregation scheduling must be completed with the lowest possible delay, so as to provide users with timely and accurate data services.
24556	但是，受信号干扰的影响，最低时延数据聚合调度问题已被证明是 NP（Non-deterministic Polynomial）难问题，而如何设计低时延的数据聚合调度算法是物联网领域的研究热点。	However, affected by signal interference, the minimum-delay data aggregation scheduling problem has been proved to be NP-hard, and how to design a low-delay data aggregation scheduling algorithm is a hot topic in the field of Internet of Things.
24557	现有面向传统物联网（如无线传感网）的解决方案通常采用逐层调度方法和干扰避免技术来实现，减少了可并发传输的链路数目，不利于降低数据聚合时延。	Most of existing solutions for traditional Internet of Things (such as wireless sensor networks) usually adopt the layer-by-layer scheduling method and the interference-avoidance technology, which is not conducive to improving data aggregation delay due to the reduced number of concurrent transmission links.
24558	值得关注的是，相继干扰消除（Successive Interference Cancellation, SIC）技术作为一种简单而强大的多包接收技术，是研究者近年来取得的重大突破，而如何结合 SIC 技术来设计物联网低时延数据聚合调度算法具有非常重要的理论研究意义。	It is worth noting that successive interference cancellation (SIC) technology, as a kind of simple and powerful multi-packet receiving technology, is a major breakthrough made by researchers in recent years. How to combine SIC technology to design low-delay data aggregation scheduling algorithms for Internet of Things has very important theoretical research significance.
24559	因此，本文以最大程度地增加可并发传输的链路数目为目标，利用跨层并发传输的思想来进行数据聚合调度，并结合SIC 技术来实现链路调度，提出了一种新颖的低时延数据聚合调度算法。	Therefore, this paper utilizes the idea of cross-layer concurrent transmission to schedule the data aggregation process, incorporates the SIC technology to schedule the data aggregation links, and proposes a novel delay-efficient data aggregation scheduling algorithm, with the aim of increasing the number of concurrent transmission links to the most extent.
24560	实验结果表明，与现有算法相比，本文所提算法在数据聚合时延优化方面最多可达43.8%。	Simulation results show that, our algorithm can improve the data aggregation delay by at most 43.8% compared with the existing algorithm.
24561	音频场景聚类的任务是将属于相同音频场景的音频样本合并到同一个类中。	Audio scene clustering (ASC) is a task to merge audio samples belonging to the same type of acoustic scene into a single cluster.
24562	本文提出一种基于联合学习框架的音频场景聚类方法。	This paper proposes a method of ASC based on joint learning framework.
24563	该框架由一个卷积自编码网络（Convolution Autoencoder Network, CAN）与一个判别性聚类网络（Discriminative Clustering Network, DCN）组成。	The proposed framework consists of a convolution autoencoder network (CAN) and a discriminative clustering network (DCN).
24564	CAN 包括编码器和译码器，用于提取深度变换特征，DCN 用于对输入的深度变换特征进行类别估计从而实现音频场景聚类。	The CAN isused to extract deep transformed feature (DTF), while the DCN is used to do cluster estimation on the input DTF for realizing ASC.