Graph Set Transformer
This work addresses the bottleneck in existing set-of-graphs architectures by fusing local and set-level information, enabling better performance on tasks requiring set-wide context.
The Graph Set Transformer (GST) introduces a neural network architecture that interleaves node-level feature propagation with cross-graph contextual modeling at every layer, outperforming baselines on synthetic and real benchmarks for set-of-graphs tasks.
We introduce the Graph Set Transformer (GST), a neural network architecture for learning on sets of graphs, designed for tasks in which per-element predictions depend on set-wide context as well as local structure. Existing architectures, including DeepSets and SetTransformer, require pre-encoded graph embeddings from a separate GNN, creating a bottleneck between feature extraction and set-level contextualisation. In contrast, GST interleaves node-level feature propagation and cross-graph contextual modelling at every layer, fusing the two levels of information through a gating mechanism. We evaluate GST on a controlled synthetic suite designed to isolate set-conditional structural reasoning and on three real-data benchmarks spanning per-atom reaction-centre identification, reaction yield prediction, and image classification. Under matched parameter budgets, GST performs better than the baselines across these settings. An architectural ablation strongly suggests that the interleaving of local and set context contributes substantially to this advantage.