Abstract

We use two sets of model parameters for GraphRNN. One larger model is used to train and test on the larger datasets that are used to compare with traditional methods. One smaller model is used to train and test on datasets with nodes up to 20. This model is only used to compare with the two most recent preliminary deep generative models for graphs proposed in (Li et al., 2018; Simonovsky & Komodakis, 2018).
For GraphRNN, the graph-level RNN uses 4 layers of GRU cells, with 128 dimensional hidden state for the larger model, and 64 dimensional hidden state for the smaller model in each layer. The edge-level RNN uses 4 layers of GRU cells, with 16 dimensional hidden state for both the larger model and the smaller model. To output the adjacency vector prediction, the edge-level RNN first maps the highest layer of the 16 dimensional hidden state to a 8 dimensional vector through a MLP with ReLU activation, then another MLP maps the vector to a scalar with sigmoid activation. The edge-level RNN is initialized by the output of the graphlevel RNN at the start of generating Sπ

Date

2018

Authors

Jiaxuan You, Rex Ying, Xiang Ren, William L Hamilton, Jure Leskovec

Journal

Proceedings of the 35th International Conference on Machine Learning, Stockholm, Sweden, PMLR

Volume

80