Yanggeng Fu

Yanggeng Fu


Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

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

Representative Publications

View All Publications
29 May 2024 in IEEE Transactions on Neural Networks and Learning Systems
Verified
Journal Article
Closed access
JLJin Li
Jin Li
QZQirong Zhang
Qirong Zhang
WLWenxi Liu
Wenxi Liu
ACAntoni B. Chan
Antoni B. Chan
YFYanggeng Fu
Yanggeng Fu
Graph neural networks (GNNs) are widely used for analyzing graph-structural data and solving graph-related tasks due to their powerful expressiveness. However, existing off-the-shelf GNN-based models usually consist of no more than three layers. Deeper GNNs usually suffer from severe performance degradation due to several issues including the infamous “over-smoothing” issue, which restricts the further development of GNNs. In this article, we investigate the over-smoothing issue in deep GNNs. We discover that over-smoothing not only results in indistinguishable embeddings of graph nodes, but also alters and even corrupts their semantic structures, dubbed semantic over-smoothing. Existing techniques, e.g., graph normalization, aim at handling the former concern, but neglect the importance of preserving the semantic structures in the spatial domain, which hinders the further improvement of model performance. To alleviate the concern, we propose a cluster-keeping sparse aggregation strategy to preserve the semantic structure of embeddings in deep GNNs (especially for spatial GNNs). Particularly, our strategy heuristically redistributes the extent of aggregations for all the nodes from layers, instead of aggregating them equally, so that it enables aggregate concise yet meaningful information for deep layers. Without any bells and whistles, it can be easily implemented as a plug-and-play structure of GNNs via weighted residual connections. Last, we analyze the over-smoothing issue on the GNNs with weighted residual structures and conduct experiments to demonstrate the performance comparable to the state-of-the-arts.

1 January 2021 in Soft Computing
Verified
Journal Article
Closed access
PDF file
NCNan-Nan Chen
Nan-Nan Chen
XGXiao-Ting Gong
Xiao-Ting Gong
YWYing-Ming Wang
Ying-Ming Wang
CZChun-Yang Zhang
Chun-Yang Zhang
YFYanggeng Fu
Yanggeng Fu

24 February 2021 in IEEE Access
Verified
Journal Article
Open access
YGYu Guan
Yu Guan
YFYanggeng Fu
Yanggeng Fu
LCLongjiang Chen
Longjiang Chen
GLGenggeng Liu
Genggeng Liu
LSLan Sun
Lan Sun
The belief-rule-base (BRB) inference methodology using the evidential reasoning (ER) approach is widely used in different fields, such as fault diagnosis, system identification, and decision analysis. However, the calculation characteristic of the conventional rule activation weight makes the inference system have the rule zero activation problem. The difficulty of constructing partial derivatives restricts the optimization of parameters using the gradient method. Hence, this paper proposes a new belief rule structure and its gradient training method to solve the rule zero activation problem during the inference process and improve inference accuracy. The Gaussian function is applied to calculate the activation weight of the rule with the new structure. Its characteristics avoid the zero activation problem caused by the attribute reference value set in the original method. Based on the newly proposed method, the corresponding distance-sensitive parameter is set for each attribute, and the weight parameter of each rule is discarded. It simplifies the calculation of rule activation weights in the inference process and enables the partial derivatives of the parameters of the inference system to be easily constructed. In the parameter optimization, the momentum optimization gradient stochastic descent method is used to train the new BRB system, which improves the training speed and accuracy compared with the conventional methods. Experiments with nonlinear function fitting, oil pipeline leak detection, and classification of several public datasets are carried out to verify whether the new BRB system trained with momentum stochastic gradient descent (SGDM-BRB) has better performance than other conventional methods. The experimental results show that in the case of complete data, SGDM-BRB has higher accuracy and faster training speed than the conventional methods.

10 October 2024 in Neural Networks
Verified
Journal Article
Closed access
YFYanggeng Fu
Yanggeng Fu
XCXinlong Chen
Xinlong Chen
SXShuling Xu
Shuling Xu
JLJin Li
Jin Li
XYXi Yao
Xi Yao
YWYing-Ming Wang
Ying-Ming Wang

27 February 2020 in IEEE Access
Verified
Journal Article
Open access
WFWeijie Fang
Weijie Fang
XGXiaoting Gong
Xiaoting Gong
GLGenggeng Liu
Genggeng Liu
YWYingjie Wu
Yingjie Wu
YFYanggeng Fu
Yanggeng Fu
The extended belief-rule-based (EBRB) system has become a widely recognized and effective rule-based system in decision-making. The system uses a data-driven method to generate the rule base by transforming each training sample into a rule. Hence, when an EBRB system is applied in an imbalanced classification dataset, the imbalance of training dataset will retain in the generated rule base. More specifically, the number of rules transformed from majority classes will be far greater than the rules transformed from minority classes. This issue usually leads to a sharp decrease in the accuracies of minority classes. This study analyses how the imbalance of training dataset exists in the generated EBRB and then proposes a Balance Adjusting (BA) approach to eliminate the influence of imbalance in the rule base. The BA approach adjusts rule activation weights of all activated rules, and further enhances the competitiveness of rules with higher activation weight during the rule aggregation process of the EBRB system. Several case studies in imbalanced benchmark classification datasets from UCI demonstrate how the use of the BA approach improves the performance of the EBRB system. This study also conducts a series of experiments to validate the improvement of the proposed approach compared with some conventional and recent existing works. The comparison results illustrate that the BA approach is feasible, effective and robust, and it performs well especially in large scale datasets. Moreover, the BA approach can also combine with various rule activation weight calculation methods, which means it might worth to be applied as a generic process before the rule aggregation process of the EBRB system.

6 March 2024 in Neural Networks
Verified
Journal Article
Closed access
QZQirong Zhang
Qirong Zhang
JLJin Li
Jin Li
QYQingqing Ye
Qingqing Ye
YLYuxi Lin
Yuxi Lin
XCXinlong Chen
Xinlong Chen
YFYanggeng Fu
Yanggeng Fu