Graph Neural Networks with convolutional ARMA filters

5 Jan 2019Filippo Maria BianchiDaniele GrattarolaCesare AlippiLorenzo Livi

Popular graph neural networks implement convolution operations on graphs based on polynomial filters defined in the spectral domain. In this paper, we propose a novel graph convolutional layer inspired by the auto-regressive moving average (ARMA) filters that, compared to the polynomial ones, are more robust and provide a more flexible graph frequency response... (read more)

PDF Abstract

Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT LEADERBOARD
Graph Regression Lipophilicity ARMA RMSE 0.894 # 8
Skeleton Based Action Recognition SBU ArmaConv Accuracy 96.00% # 4