Recurrent Temporal Revision Graph Networks

TL;DR

This paper introduces a recurrent neural network framework for temporal neighbor aggregation in temporal graphs, effectively capturing complete historical neighbor information and outperforming existing methods in real-world applications.

Contribution

It proposes a novel recurrent neural network-based framework for temporal neighbor aggregation that addresses bias and incompleteness in existing subsampling methods.

Findings

Achieves +9.6% improvement in averaged precision on Ecommerce dataset.

Demonstrates superior theoretical expressiveness.

Outperforms state-of-the-art methods in real-world tasks.

Abstract

Temporal graphs offer more accurate modeling of many real-world scenarios than static graphs. However, neighbor aggregation, a critical building block of graph networks, for temporal graphs, is currently straightforwardly extended from that of static graphs. It can be computationally expensive when involving all historical neighbors during such aggregation. In practice, typically only a subset of the most recent neighbors are involved. However, such subsampling leads to incomplete and biased neighbor information. To address this limitation, we propose a novel framework for temporal neighbor aggregation that uses the recurrent neural network with node-wise hidden states to integrate information from all historical neighbors for each node to acquire the complete neighbor information. We demonstrate the superior theoretical expressiveness of the proposed framework as well as its state-of-the-art performance in real-world applications. Notably, it achieves a significant +9.6% improvement on averaged precision in a real-world Ecommerce dataset over existing methods on 2-layer models.