DFNets: Spectral CNNs for Graphs with Feedback-Looped Filters

TL;DR

DFNets introduce a spectral CNN model with feedback-looped filters for graph data, achieving better localization, convergence guarantees, and superior performance on classification benchmarks.

Contribution

The paper presents a novel spectral CNN architecture with feedback-looped filters that improve localization, convergence, and applicability to any graph structure.

Findings

Outperforms state-of-the-art methods on citation network classification.

Achieves superior results on knowledge graph entity classification.

Provides theoretical guarantees for filter convergence.

Abstract

We propose a novel spectral convolutional neural network (CNN) model on graph structured data, namely Distributed Feedback-Looped Networks (DFNets). This model is incorporated with a robust class of spectral graph filters, called feedback-looped filters, to provide better localization on vertices, while still attaining fast convergence and linear memory requirements. Theoretically, feedback-looped filters can guarantee convergence w.r.t. a specified error bound, and be applied universally to any graph without knowing its structure. Furthermore, the propagation rule of this model can diversify features from the preceding layers to produce strong gradient flows. We have evaluated our model using two benchmark tasks: semi-supervised document classification on citation networks and semi-supervised entity classification on a knowledge graph. The experimental results show that our model considerably outperforms the state-of-the-art methods in both benchmark tasks over all datasets.