Abstract:In recent years, graph neural network (GNN) has become the main stream for most of recent researches due to its powers in dealing with complex graph data learning problems. However, as most of the recent GNN-based architectures have been mainly designed to only evaluate direct relational structures between nodes. As the results, these techniques are unable to capture the sophisticated multi-way relationships in graph. The multi-way relationships can be represented by both explicit graph-based and complex topological structures. In general, the multi-way relationships in graph can be modelled as simplicial complexes, hyper-graphs, e.g., and can be efficiently preserved under the simplicial neural networks (SNN). There are several notable SNN-based architectures have been proposed recently, such as the well-known simplicial convolutional neural network (SCNN). The SNN-based techniques have shown the competitive performances in handling graph learning. However, most of recent proposed SNN-based architectures are designed upon the deep neural learning paradigm. Therefore, they still encountered several challenges with regard to the feature noise and data uncertainty. To overcome these limitations, in this paper, we proposed a novel integrated SNN and neuro-fuzzy network (NFN) graph embedding technique, called as: SFGE. Our SFGE model is designed to better capture the multi-way structural representation of graph by taking advances of different advanced graph-based and fuzzy-based neural learning techniques. By taking advances of neuro-fuzzy learning approach, our model can efficiently support to eliminate the feature uncertainty/ambiguity during the task-driven fine-tuning process. In addition, it also supports to better capture the rich multi-way relational structures of the input graphs under the topology-enhanced graph analysis approach. Extensive empirical studies within a real-world molecular graph dataset have effectiveness of our SFGE in dealing with graph classification task.

Genetically Optimized Multi-Layer Fuzzy Polynomial Neural Networks: Analysis and Design.

Fuzzy reinforced polynomial neural networks constructed with the aid of PNN architecture and fuzzy hybrid predictor based on nonlinear function

The learning algorithm for a novel fuzzy neural network

Multiobjective Evolution of the Explainable Fuzzy Rough Neural Network with Gene Expression Programming

Optimal Method of Neuro-Fuzzy Network and Its Application Based on Variable Precision Rough Set

Relationship Between a Fuzzy Logic and a Steepest Descent Approach to Optimize a Feedforward Artificial Neural Network Configuration.

Multi-Objective Fuzzy Modeling Using Nsga-Ii

Random Polynomial Neural Networks: Analysis and Design

Fuzzy Neural Network Optimization Method Based on Hopfield Networks

Graph Fuzzy System: Concepts, Models and Algorithms

Disjunctive Fuzzy Neural Networks: A New Splitting-Based Approach to Designing a T–S Fuzzy Model

Deep Fuzzy Min–Max Neural Network: Analysis and Design

OPTIMUM DESIGN OF MLP NETWORK STRUCTURE USING FUZZY LOGIC APPROACH

An integrated simplicial neural network with neuro-fuzzy network for graph embedding

Fuzzy Neural Network Model For A Class Of Nonlinear Systems

Neural Networks Designing Neural Networks: Multi-Objective Hyper-Parameter Optimization

A development of granular logic neural networks

An Optimal Design Approach for Fuzzy Systems Based on Hybrid Genetic Algorithms

Messy Genetic Algorithm Based New Learning Method for Structurally Optimised Neurofuzzy Controllers

Neural Network Based Fuzzy Identification and Its Application to Modeling and Control of Complex Systems

An Efficient Self-Organizing Deep Fuzzy Neural Network for Nonlinear System Modeling