LGAIMLJun 22, 2022

Agent-based Graph Neural Networks

ETH Zurich
arXiv:2206.11010v228 citationsh-index: 81
Originality Highly original
AI Analysis

This work addresses graph classification challenges for machine learning researchers, offering a novel approach with theoretical guarantees, though it is incremental in advancing GNN architectures.

The authors tackled the problem of graph-level tasks by introducing AgentNet, a graph neural network with computational complexity independent of graph size, which outperformed standard and more expensive GNNs in synthetic and real-world experiments.

We present a novel graph neural network we call AgentNet, which is designed specifically for graph-level tasks. AgentNet is inspired by sublinear algorithms, featuring a computational complexity that is independent of the graph size. The architecture of AgentNet differs fundamentally from the architectures of traditional graph neural networks. In AgentNet, some trained \textit{neural agents} intelligently walk the graph, and then collectively decide on the output. We provide an extensive theoretical analysis of AgentNet: We show that the agents can learn to systematically explore their neighborhood and that AgentNet can distinguish some structures that are even indistinguishable by 2-WL. Moreover, AgentNet is able to separate any two graphs which are sufficiently different in terms of subgraphs. We confirm these theoretical results with synthetic experiments on hard-to-distinguish graphs and real-world graph classification tasks. In both cases, we compare favorably not only to standard GNNs but also to computationally more expensive GNN extensions.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes