Abstract
AbstractGraph neural networks (GNNs), which generalize deep neural network models to graph structure data, have attracted increasing attention and achieved state-of-the-art performance in graph-related tasks such as graph classification, link prediction, and node classification. To adapt GNNs to graph classification, existing works aim to define the graph pooling method to learn graph-level representation by downsampling and summarizing the information present in the nodes. However, most existing pooling methods lack a way of obtaining information about the entire graph from both the local and global aspects of the graph. Moreover, in these pooling methods, the difference features between nodes and their neighbors are usually ignored, which is crucial in obtaining graph information in our opinions. In this paper, we propose a novel graph pooling method called Node Information Awareness Pooling (NIAPool), which addresses the limitations of previous graph pooling methods. NIAPool utilizes a novel self-attention framework and a new convolution operation that can better capture the difference features between nodes to obtain node information in the graph from both local and global aspects. Experiments on five public benchmark datasets demonstrate the superior performance of NIAPool for graph classification compared to the state-of-the-art baseline methods.KeywordsGraph poolingGraph classificationGraph representation learningGraph neural network
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.