DATA-DRIVEN LEARNING OF GEOMETRIC SCATTERING MODULES FOR GNNS.
DATA-DRIVEN LEARNING OF GEOMETRIC SCATTERING MODULES FOR GNNS.
复制标题
DOI:
10.1109/mlsp52302.2021.9596169
复制
发表时间:
2021-10
期刊:
影响因子:
--
通讯作者:
Wolf, Guy
中科院分区:
文献类型:
--
作者:
Tong, Alexander;Wenkel, Frederick;Macdonald, Kincaid;Krishnaswamy, Smita;Wolf, Guy
We propose a new graph neural network (GNN) module, based on relaxations of recently proposed geometric scattering transforms, which consist of a cascade of graph wavelet filters. Our learnable geometric scattering (LEGS) module enables adaptive tuning of the wavelets to encourage band-pass features to emerge in learned representations. The incorporation of our LEGS-module in GNNs enables the learning of longer-range graph relations compared to many popular GNNs, which often rely on encoding graph structure via smoothness or similarity between neighbors. Further, its wavelet priors result in simplified architectures with significantly fewer learned parameters compared to competing GNNs. We demonstrate the predictive performance of LEGS-based networks on graph classification benchmarks, as well as the descriptive quality of their learned features in biochemical graph data exploration tasks.
登录
查看更多内容
影响因子:
2.5
作者:
Coifman, Ronald R.;Maggioni, Mauro
通讯作者:
Maggioni, Mauro
影响因子:
2.9
作者:
Ovchinnikov S;Park H;Kim DE;DiMaio F;Baker D
通讯作者:
Baker D
影响因子:
2.9
作者:
Modi, Vivek;Xu, Qifang;Dunbrack, Roland L., Jr.
通讯作者:
Dunbrack, Roland L., Jr.
影响因子:
2.9
作者:
Moult J;Fidelis K;Kryshtafovych A;Schwede T;Tramontano A
通讯作者:
Tramontano A
影响因子:
3
作者:
Mallat, Stephane
通讯作者:
Mallat, Stephane