On the Equivalence Between Temporal and Static Graph Representations for Observational Predictions
In this work we formalize the (pure observational) task of predicting node attribute evolution in temporal graphs. We show that node representations of temporal graphs can be cast into two distinct frameworks: (a) The de-facto standard approach, which we denote time-and-graph, where equivariant graph (e.g., GNN) and sequence (e.g., RNN) representations are intertwined to represent the temporal evolution of the graph; and (b) an approach that we denote time-then-graph, where the sequences describing the node and edge dynamics are represented first (e.g., RNN), then fed as node and edge attributes into a (static) equivariant graph representation that comes after (e.g., GNN). In real-world datasets, we show that our time-then-graph framework achieves the same prediction performance as state-of-the-art time-and-graph methods. Interestingly, time-then-graph representations have an expressiveness advantage over time-and-graph representations when both use component GNNs that are not most-expressive (e.g., 1-Weisfeiler-Lehman GNNs). We introduce a task where this expressiveness advantage allows time-then-graph methods to succeed while state-of-the-art time-and-graph methods fail.
READ FULL TEXT