Diffusion-Convolutional Neural Networks
James Atwood, Don Towsley
Code Available — Be the first to reproduce this paper.
ReproduceCode
Abstract
We present diffusion-convolutional neural networks (DCNNs), a new model for graph-structured data. Through the introduction of a diffusion-convolution operation, we show how diffusion-based representations can be learned from graph-structured data and used as an effective basis for node classification. DCNNs have several attractive qualities, including a latent representation for graphical data that is invariant under isomorphism, as well as polynomial-time prediction and learning that can be represented as tensor operations and efficiently implemented on the GPU. Through several experiments with real structured datasets, we demonstrate that DCNNs are able to outperform probabilistic relational models and kernel-on-graph methods at relational node classification tasks.
Tasks
Benchmark Results
| Dataset | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| CiteSeer (0.5%) | DCNN | Accuracy | 53.1 | — | Unverified |
| CiteSeer (1%) | DCNN | Accuracy | 62.2 | — | Unverified |
| CiteSeer with Public Split: fixed 20 nodes per class | DCNN | Accuracy | 69.4 | — | Unverified |
| Cora (0.5%) | DCNN | Accuracy | 59 | — | Unverified |
| Cora (1%) | DCNN | Accuracy | 66.4 | — | Unverified |
| Cora (3%) | DCNN | Accuracy | 76.7 | — | Unverified |
| Cora with Public Split: fixed 20 nodes per class | DCNN | Accuracy | 79.7 | — | Unverified |
| PubMed (0.03%) | DCNN | Accuracy | 60.9 | — | Unverified |
| PubMed (0.05%) | DCNN | Accuracy | 66.7 | — | Unverified |
| PubMed (0.1%) | DCNN | Accuracy | 73.1 | — | Unverified |
| PubMed with Public Split: fixed 20 nodes per class | DCNN | Accuracy | 76.8 | — | Unverified |