Temporal Generalization Estimation in Evolving Graphs
This addresses the challenge of temporal generalization for practitioners using GNNs in evolving real-world graphs, though it is incremental as it builds on existing methods with a novel adaptation.
The paper tackles the problem of Graph Neural Networks (GNNs) struggling to maintain accurate representations as graphs evolve, by introducing Smart, a method that uses self-supervised graph reconstruction to estimate temporal distortion, achieving a MAPE of 2.19% on the OGB-arXiv dataset.
Graph Neural Networks (GNNs) are widely deployed in vast fields, but they often struggle to maintain accurate representations as graphs evolve. We theoretically establish a lower bound, proving that under mild conditions, representation distortion inevitably occurs over time. To estimate the temporal distortion without human annotation after deployment, one naive approach is to pre-train a recurrent model (e.g., RNN) before deployment and use this model afterwards, but the estimation is far from satisfactory. In this paper, we analyze the representation distortion from an information theory perspective, and attribute it primarily to inaccurate feature extraction during evolution. Consequently, we introduce Smart, a straightforward and effective baseline enhanced by an adaptive feature extractor through self-supervised graph reconstruction. In synthetic random graphs, we further refine the former lower bound to show the inevitable distortion over time and empirically observe that Smart achieves good estimation performance. Moreover, we observe that Smart consistently shows outstanding generalization estimation on four real-world evolving graphs. The ablation studies underscore the necessity of graph reconstruction. For example, on OGB-arXiv dataset, the estimation metric MAPE deteriorates from 2.19% to 8.00% without reconstruction.