Pathfinder Discovery Networks for Neural Message Passing
Authors:
Open Access
- 19 April 2021
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 2547-2558
Abstract
In this work we propose Pathfinder Discovery Networks (PDNs), a method for jointly learning a message passing graph over a multiplex network with a downstream semi-supervised model. PDNs inductively learn an aggregated weight for each edge, optimized to produce the best outcome for the downstream learning task. PDNs are a generalization of attention mechanisms on graphs which allow flexible construction of similarity functions between nodes. They also support edge convolutions and cheap multiscale mixing layers. We show that PDNs overcome weaknesses of existing methods for graph attention (e.g. Graph Attention Networks), such as the diminishing weight problem. Our experimental results demonstrate competitive predictive performance on academic node classification tasks. Additional results from a challenging suite of node classification experiments show how PDNs can learn a wider class of functions than existing baselines. We analyze the relative computational complexity of PDNs, and show that PDN runtime is not considerably higher than static-graph models. Finally, we discuss how PDNs can be used to construct an easily interpretable attention mechanism that allows users to understand information propagation in the graph.Keywords
All Related Versions
This publication has 27 references indexed in Scilit:
- Don't Walk, Skip!2017
- node2vec2016
- Asymmetric Transitivity Preserving Graph Embedding2016
- GraRep2015
- LINE2015
- DeepWalk2014
- Weighted Multiplex NetworksPLOS ONE, 2014
- Influence of Graph Construction on Semi-supervised Learning2013
- Predicting missing links via local informationThe European Physical Journal B, 2009
- Friends and neighbors on the WebSocial Networks, 2003