ID 原文 译文
57918 设计了 6 个探测任务，覆盖句子浅层、语法和语义三方面信息. Six probing tasks are given to cover the shallow information of sentences，grammar and semantics.
57919 最后在 6 个文本分类数据集上验证了本文的方法，使分类错误率得到改善. The method is shown validated on six text classification datasets，and classification error rate is improved.
57920 智能人机对话系统综合应用了人工智能领域多项核心技术，随着深度学习、强化学习等基础算法的快速发展，人机对话系统的总体架构、算法体系以及应用模式都有了重大的改变和提升. Intelligent human-machine dialogue system comprehensively utilizes a number of core technol- ogies in the field of artificial intelligence. In recent years，with the development of basic algorithms such as deep learning and reinforcement learning，the overall structure，algorithm system and application mode of human-machine dialogue system have been changed and improved greatly.
57921 从人机对话系统的架构、人机对话系统的核心算法和该领域面临的挑战以及技术研究方面做了综述. In order to sort out and summarize this technology，the architecture of human-machine dialogue system，the core algorithm of hu- man-machine dialogue system，challenges and technical directions in this field are reviewed.
57922 针对基于媒介调制的广义空间调制( GSM-MBM) 系统接收端最大似然( ML) 检测算法计算复杂度高且随激活天线数呈指数递增的问题，提出一种基于能量排序下的松弛迭代思想的低复杂度检测算法( EO-ＲIM) . The complexity of the maximum likelihood ( ML ) detector of the generalized spatialmodulation-media based modulation system is very high and exponentially grows with the number of activeantennas. A low-complexity detection algorithm termed energy ordered-relaxation iteration method ( EO-ＲIM) is proposed.
57923 该算法对所有可能的发射天线组合及相应镜像激活模式组合下的信号能量总值进行排序，再通过松弛迭代算法依次检测相应的调制信号，并通过预设阈值来协调误码率( BEＲ) 性能和计算复杂度之间的关系. First，the possible active transmit antenna combinations and corresponding mirroractivation pattern combinations are sorted according to their signal energy，then a relaxation iterativemethod is performed to obtain corresponding modulated signals.
57924 仿真结果表明，在 GSMMBM 系统中，EO-ＲIM 算法的 BEＲ 性能逼近 ML 检测算法，与基于有序块的最小均方误差( OB-MMSE) 检测算法几乎一致，而 EO-ＲIM 的计算复杂度随激活天线数呈平方递增而非指数递增，相比 OB-MMSE 算法降低了一个数量级. According to a predefined threshold，thealgorithm strikes a trade-off between complexity and performance. Simulations show that the bit error ratioperformance of EO-ＲIM algorithm approaches that of ML detection algorithm and is comparable to that ofthe ordered block minimum mean squared error detection algorithm. The computational complexity of EO-ＲIM grows with the square of the number of active antennas，while ML detector has exponentialcomplexity.
57925 提出了一种面向超密集场景的考虑业务动态的无线网络功率资源匹配算法. A wireless network resource matching algorithm considering service dynamics in ultra-densescenarios is proposed.
57926 首先，根据网络异构和业务动态变化特征，建立双层动态博弈模型. Firstly，according to heterogeneous characteristics of networks and dynamic changes of services，a two-layer dynamic game model was established.
57927 针对不同博弈层参与者的需求特性，以最大化效用函数为准则，设计不同的效益模型，并通过对网络中业务动态性的预测调整定价因子，以更准确地反映网络环境的变化; In particular，different benefit modelswere designed to maximize the utility function based on the demand characteristics of different game layerparticipants，and a pricing factor was dynamically adjusted by predicting service dynamics more accurately to reflect the network environment.