孙忆南 徐雅静 李思 郭军
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0005
摘要 ( 607 ) PDF (1628 KB)( 40 )
Proactive dialogue generates dialogue utterance based on a conversation goal and a given knowledge graph (KG). Existing methods combine knowledge of each turn of dialogue with knowledge triples by hidden variables, resulting in the interpretability of generation results is relatively poor. An interpretable knowledge-aware path (KAP) model was proposed for knowledge reasoning in proactive dialogue generation. KAP model can transform explicit and implicit knowledge of each turn of dialogue into corresponding dialogue state matrix, thus forming the KAP for dialogue history. Based on KAP, the next turn of dialogue state vector can be infered from both the topology and semantic of KG. This vector can indicate knowledge distribution of next sentence, so it enhances the accuracy and interpretability of dialogue generation. Experiments show that KAP model’s dialogue generation is closer to actual conversation than other state-of-the-art proactive dialogue models.
陈彦萍 靳晓东 夏虹 王忠民
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0003
摘要 ( 462 ) PDF (5325 KB)( 49 )
Industrial big data was usually multi-source, heterogeneous, and deeply intertwined. It had a wide range of data sources, high data dimensions, and strong data correlation. In order to effectively analyze and process streaming industrial big data generated by edge computing, it was very important to provide an effective real-time incremental data method. However, in the process of incremental processing, industrial big data incremental computing faced the challenges of dimensional disaster, repeated calculations, and the explosion of intermediate results. Therefore, in order to solve the above problems effectively, a QR-based tensor-train (TT) decomposition (TTD) method and a QR-based incremental TTD (QRITTD) method were proposed. This algorithm combined the incremental QR-based decomposition algorithm with an approximate singular value decomposition ( SVD) algorithm and had good scalability. In addition, the computational complexity, space complexity, and approximation error analysis were analyzed in detail. The effectiveness of the three algorithms of QRITTD, non-incremental TTD (NITTD), and TT rank-1 (TTr1) SVD (TTr1SVD)were verified by comparison. Experimental results show that the SVD QRITTD method has better performance under the premise of ensuring the same tensor size.
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0013
摘要 ( 207 ) PDF (7836 KB)( 90 )
To improve the anti-noise ability of fuzzy local information C-means clustering, a robust entropy-like distance driven fuzzy clustering with local information is proposed. This paper firstly uses Jensen-Shannon divergence to induce a symmetric entropy-like divergence. Then the root of entropy-like divergence is proved to be a distance measure, and it is applied to existing fuzzy C-means (FCM) clustering to obtain a new entropy-like divergence driven fuzzy clustering, meanwhile its convergence is strictly proved by Zangwill theorem. In the end, a robust fuzzy clustering by combing local information with entropy-like distance is constructed to segment image with noise. Experimental results show that the proposed algorithm has better segmentation accuracy and robustness against noise than existing state-of-the-art fuzzy clustering-related segmentation algorithm in the presence of noise.
韩凤全 韩英华 陆靖 赵强
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0004
摘要 ( 266 ) PDF (4122 KB)( 93 )
With the expansion of wind speed data sets, decreasing model training time is of great significance to the time cost of wind speed prediction. And imperfection of the model evaluation system also affect the wind speed prediction. To address these challenges, a hybrid method based on feature extraction, nested shared weight long short-term memory (NSWLSTM) network and Gaussian process regression (GPR) was proposed. The feature extraction of wind speed promises the best performance of the model. NSWLSTM model reduces the training time of long short-term memory (LSTM) network and improves the prediction accuracy. Besides, it adopted a method combined NSWLSTM with GPR (NSWLSTMGPR) to provide the probabilistic prediction of wind speed. The probabilistic prediction can provide information that deviates from the predicted value, which is conducive to risk assessment and optimal scheduling. The simulation results show that the proposed method can obtain high-precision point prediction, appropriate prediction interval and reliable probabilistic prediction results with shorter training time on the wind speed prediction.
刘君 胡予濮 陈杰
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0006
Digital rights management (DRM) applications are usually confronted with threats like key extraction, code lifting, and illegal distribution. White-box cryptography aims at protecting software implementations of cryptographic algorithms and can be employed into DRM applications to provide security. A general DRM solution based on white-box cryptography was proposed to address the three threats mentioned above. The method is to construct a general perturbation-enabled white-box compiler for lookup-table based white-box block ciphers, such that the white-box program generated by this compiler provides traceability along with resistance against key extraction and code lifting. To get a traceable white-box program, the idea of hiding a slight perturbation in the lookup-table was employed, aiming at perturbing its decryption functionality, so that each user can be identified. Security analysis and experimental results show that the proposed DRM solution is secure and practical.
赵国生 张婧婷 王健
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0002
摘要 ( 248 ) PDF (3819 KB)( 46 )
In view of the privacy security issues such as location information leakage in the interaction process between the base station and the sensor nodes in the sensor-cloud system, a base station location privacy protection algorithm based on local differential privacy (LDP) is proposed. Firstly, through the local obfuscation algorithm (LOA), the base station can get the data of the real location and the pseudo location by flipping a coin, and then send the data to the fog layer, then the obfuscation location domain set is obtained. Secondly, in order to reconstruct the location distribution of the real location and the pseudo location in the base station, the location domain of the base station is divided into several decentralized sub-regions, and a privacy location reconstruction algorithm ( PLRA) is performed in each sub-region. Finally, the base station correlates the location information of each sub-region, and
then uploads the data information containing the disturbance location to the fog node layer. The simulation results show that compared with the existing base station location anonymity and security technique (BLAST) algorithm, the proposed method not only reduce the algorithm's running time and network delay, but also improve the data availability. So the proposed method can protect the location privacy of the base station more safely and efficiently.
陈嘉宁 李曦 纪红 张鹤立
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0001
摘要 ( 249 ) PDF (2262 KB)( 75 )
Pushing popular contents to the edge of the network can meet the growing demand for data traffic, reduce latency and relieve the pressure of the backhaul. However, considering the limited storage space of the base stations, it is impossible to cache all the contents, especially in ultra-dense network ( UDN). Furthermore, the uneven distribution of mobile users results in load imbalance among small base stations (SBSs) in both time and space, which also affects the caching strategy. To overcome these shortcoming, the impact of the changing load imbalance in UDN was investigated, and then a dynamic hierarchical collaborative caching (DHCC) scheme was proposed to optimize latency and caching hit rate. The storage of the SBS is logically divided into the independent caching layer and the collaborative caching layer. The independent caching layer caches the most popular contents for local users爷interest, and the collaborative caching layer caches contents as much as possible for the benefit of content diversity in the region. Different SBSs have respective storage space layer division ratios, according to their real-time traffic load. For SBSs with heavy load, the independent caching layers are allocated with more space. Otherwise, the collaborative caching layers could store more contents with larger space. The simulation results show that, DHCC improved both transmission latency and hit rate compared with existing caching schemes.
敦书波 金秋延 毕凌宇 王紫任 谢军
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0007
摘要 ( 230 ) PDF (3245 KB)( 28 )
Coaxial connectors are generally regarded as a kind of potential passive non-linear source when magnetic materials are applied in the coating or under-plating, which may result in serious passive intermodulation (PIM) interference and degrade the communication quality. In this paper, the effect of connector coating materials on the PIM is theoretically studied using finite element analysis (FEA) and circuit simulations. Considering the material composition both in central and outer conductor, an FEA model of connector is proposed to identify the current density in magnetic material region. An equivalent circuit model expressing the nonlinearity in coating material is developed, coupled with the non-linear transfer model. The PIM product power of the connector with related material configuration is predicted by harmonic balance simulation. Intentionally design connector samples are used in PIM tests and the measurement results are consistent with the theoretical predictions. The PIM performance in coaxial connectors is demonstrated from the perspectives of both modeling analysis and experimental investigations.
邓军勇 吴皓月 罗佳莹 山蕊 刘新闯
中国邮电高校学报(英文版), 2021, 28 (1). doi： 10.19682/j.cnki.1005-8885.2021.0012
摘要 ( 270 ) PDF (5381 KB)( 43 )
Low-loss, non-blocking, scalable passive optical interconnect network on-lhip ( LOOKNoC) structure was proposed based on 2 *2 optical exchange switches, using wavelength division multiplexing (WDM) technology to expand to 8 *8, 16 *16, 32 *32, 64 *64 passive optical interconnection networks, which can achieve non-blocking communication. The experimental results show that based on the 16 *16 optical interconnection network structure, the number of microring resonators (MRs) in LOOKNoC was reduced by 90.9%, 90.9%, 20.0% and 75.0% compared with the generic wavelength-routed optical router (GWOR), λ-router topology and CrossBar structure. By testing the performance parameters based on the structure of 16 *16 by the OMNET + + platform, as the result shows, the average insertion loss of LOOKNoC is 3.0%, 11.6%, 4.8% and 16.7% less than that of
GWOR, λ-router, Mesh, and CrossBar structures.