Fast Graph Kernel with Optical Random Features

TL;DR

This paper introduces a fast graph kernel method using optical random features, significantly reducing computation time while maintaining or improving classification accuracy in graph analysis.

Contribution

It proposes integrating optical random features into the graphlet kernel framework, establishing a theoretical link with mean kernel metrics, and demonstrating substantial speed improvements.

Findings

Orders of magnitude faster than traditional graphlet kernel

Maintains or improves classification accuracy

Theoretical link with mean kernel metric

Abstract

The graphlet kernel is a classical method in graph classification. It however suffers from a high computation cost due to the isomorphism test it includes. As a generic proxy, and in general at the cost of losing some information, this test can be efficiently replaced by a user-defined mapping that computes various graph characteristics. In this paper, we propose to leverage kernel random features within the graphlet framework, and establish a theoretical link with a mean kernel metric. If this method can still be prohibitively costly for usual random features, we then incorporate optical random features that can be computed in constant time. Experiments show that the resulting algorithm is orders of magnitude faster that the graphlet kernel for the same, or better, accuracy.