Jun Zeng

Jun Zeng

Auto-generated profile, based on publicly available information

Jun Zeng

College of Telecommunications and Information Engineering, Nanjing University of Posts and Telecommunications, Nanjing, China

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 May 2021 in IEEE Transactions on Cognitive Communications and Networking
Verified
Journal Article
Closed access
JZJun Zeng
Jun Zeng
JSJinlong Sun
Jinlong Sun
GGGuan Gui
Guan Gui
Bamidele AdebisiBamidele Adebisi
Bamidele Adebisi
Tomoaki OhtsukiTomoaki Ohtsuki
Tomoaki Ohtsuki
HSHikmet Sari
Hikmet Sari
In this paper, a channel state information (CSI) feedback method is proposed based on deep transfer learning (DTL). The proposed method addresses the problem of high training cost of downlink CSI feedback network in frequency division duplexing (FDD) massive multiple-input multiple-output (MIMO) systems. In particular, we obtain the models of different wireless channel environments at low training cost by fine-tuning the pre-trained model with a relatively small number of samples. In addition, the effects of different layers on training cost and model performance are discussed. Furthermore, a model-agnostic meta-learning (MAML)-based method is proposed to solve the problem associated with large number of samples of a wireless channel environment required to train a deep neural network (DNN) as a pre-trained model. Our results show that the performance of the DTL-based method is comparable with that of the DNN trained with a large number of samples, which demonstrates the effectiveness and superiority of the proposed method. At the same time, although there is a certain performance loss compared with the DTL-based method, the MAML-based method shows good performance in terms of the normalized mean square error (NMSE).

12 November 2019 in IEEE Access
Verified
Journal Article
Open access
JZJun Zeng
Jun Zeng
JWJuan Wang
Juan Wang
LGLiang Guo
Liang Guo
GFGuanghui Fan
Guanghui Fan
KZKaixuan Zhang
Kaixuan Zhang
GGGuan Gui
Guan Gui

18 November 2019 in IEEE Access
Verified
Journal Article
Open access
GFGuanghui Fan
Guanghui Fan
WSWenjuan Shi
Wenjuan Shi
LGLiang Guo
Liang Guo
JZJun Zeng
Jun Zeng
KZKaixuan Zhang
Kaixuan Zhang
GGGuan Gui
Guan Gui
Cellular network performance is often evaluated by key performance indicator (KPI) and key quality indicator (KQI). The association between KQI and KPI is the most critical step to optimize the performance of cellular network. Traditional association methods between KPI and KQI are based on the end-to-end evaluation. However, these methods require professional engineers to evaluate cellular network performance, and the man-manned evaluation is often inaccurate and labor-consuming, which cannot find out the caused factor of deterioration networks. In order to solve the problem, we propose a machine learning based quantitative association rule mining (QARM) method called SWP-RF, which consists of sliding-window partitioning (SWP) and random forest (RF), to associate KPI with KQI. Specifically, we first use SWP to discretize continuous attributes into boolean values, which is adopted to mine its association rules. The warning intervals and the warning points is obtained by SWP. Next, we use RF feature importance to measure the association between KPI and KQI. The priority of the association strength between all KPIs and each KQI can be obtained by RF. Finally, we select the warning points and the association strength priority as optimal output solution. Experiments are conducted over the actual data from telecom operators and the results confirm the feasibility and accuracy of the proposed method.

27 November 2019 in IEEE Access
Verified
Journal Article
Open access
KZKaixuan Zhang
Kaixuan Zhang
JWJuan Wang
Juan Wang
WZWei Zhang
Wei Zhang
KWKe Wang
Ke Wang
JZJun Zeng
Jun Zeng
GGGuan Gui
Guan Gui
The quality of wireless user perception for cells in a particular scenario is reflected on a set of indicators. Comprehensive evaluation of those cells is the base of network optimization for operators. Traditional methods use weighted sum of all indicators as the evaluation result. However, these indicators include some ineffective ones, which leads to an unconvincing evaluation result. To achieve a convincing and accurate result, we propose a lightweight comprehensive evaluation method. Firstly, indicator selection is implemented via random forest algorithm. Secondly, those selected indicators are weighted via entropy method. Finally, we compute the score of all cells with the weights. Experiment results are given to show that the cells with higher score perform better on all indicators, which is coincide with the actual situation. Hence, our proposed method is not only lightweight but also obtain more accurate result.