REVIEW 4 cited by
Do Transformers Really Perform Bad for Graph Representation?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
The Transformer architecture has become a dominant choice in many domains, such as natural language processing and computer vision. Yet, it has not achieved competitive performance on popular leaderboards of graph-level prediction compared to mainstream GNN variants. Therefore, it remains a mystery how Transformers could perform well for graph representation learning. In this paper, we solve this mystery by presenting Graphormer, which is built upon the standard Transformer architecture, and could attain excellent results on a broad range of graph representation learning tasks, especially on the recent OGB Large-Scale Challenge. Our key insight to utilizing Transformer in the graph is the necessity of effectively encoding the structural information of a graph into the model. To this end, we propose several simple yet effective structural encoding methods to help Graphormer better model graph-structured data. Besides, we mathematically characterize the expressive power of Graphormer and exhibit that with our ways of encoding the structural information of graphs, many popular GNN variants could be covered as the special cases of Graphormer.
Forward citations
Cited by 4 Pith papers
-
Graph Neural Networks for the Graphical Bootstrap
GNNs and graph transformers classify vanishing coefficients on millions of N=4 SYM f-graphs, generalizing to larger n with 99.996% ROC AUC and pruning up to 85.5% of redundant d-graphs.
-
ReDiSC: A Reparameterized Masked Diffusion Model for Scalable Node Classification with Structured Predictions
A reparameterized masked diffusion model with variational EM gives scalable structured node classification, matching or beating GNN, label propagation, and continuous diffusion baselines.
-
Polaritonic Machine Learning for Graph-based Data Analysis
Simulated polariton condensate lattices act as physics-based feature generators for CNNs and improve classification of cliques and asymmetries in point clouds over raw point images in three synthetic tasks.
-
QLPO: Quadrant-weighted Sampling for Length-aware Policy Optimization
QLPO resamples GRPO training groups to favor short correct and long incorrect responses, cutting reasoning length substantially while keeping accuracy roughly unchanged.
Discussion (0). Continue with ORCID to comment.