摘要: Graph classification is a problem with practical applications in many different domains. Most of the existing methods take entire graph into account when calculating features. In graphlet-based approach, for instance, processed to get total count graphlets or sub-graphs. real-world, however, graphs can be both large and noisy discriminative patterns confined certain regions only. this work, we study attentional processing classification. The use attention allows us focus on small but informative parts graph, avoiding noise rest graph. We present novel RNN model, called Attention Model (GAM), that processes only portion by adaptively selecting sequence "interesting" nodes. model equipped an external memory component which it integrate information gathered from demonstrate effectiveness through various experiments.
arxiv.org
harvard.edu参考文章(4)
Ronald J. Williams, Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning Machine Learning. ,vol. 8, pp. 229- 256 ,(1992) , 10.1007/BF00992696
Ram Nevatia, Liang-Chieh Chen, Kan Chen, Haoyuan Gao, Jiang Wang, Wei Xu, ABC-CNN: An Attention Based Convolutional Neural Network for Visual Question Answering arXiv: Computer Vision and Pattern Recognition. ,(2015)
Gabriel Dulac-Arnold, Peter Sunehag, Ben Coppin, Richard Evans, Reinforcement Learning in Large Discrete Action Spaces. arXiv: Artificial Intelligence. ,(2015)
Christopher D. Manning, Minh-Thang Luong, Hieu Pham, Effective Approaches to Attention-based Neural Machine Translation arXiv: Computation and Language. ,(2015)