REVIEW 8 cited by
A Fair Comparison of Graph Neural Networks for Graph Classification
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
Experimental reproducibility and replicability are critical topics in machine learning. Authors have often raised concerns about their lack in scientific publications to improve the quality of the field. Recently, the graph representation learning field has attracted the attention of a wide research community, which resulted in a large stream of works. As such, several Graph Neural Network models have been developed to effectively tackle graph classification. However, experimental procedures often lack rigorousness and are hardly reproducible. Motivated by this, we provide an overview of common practices that should be avoided to fairly compare with the state of the art. To counter this troubling trend, we ran more than 47000 experiments in a controlled and uniform framework to re-evaluate five popular models across nine common benchmarks. Moreover, by comparing GNNs with structure-agnostic baselines we provide convincing evidence that, on some datasets, structural information has not been exploited yet. We believe that this work can contribute to the development of the graph learning field, by providing a much needed grounding for rigorous evaluations of graph classification models.
Forward citations
Cited by 8 Pith papers
-
Bridging Theory and Practice in Link Representation with Graph Neural Networks
A new framework classifies message-passing link representation models by neighborhood radius and base expressiveness, yielding a hierarchy in which SEAL is most expressive, plus a synthetic benchmark and symmetry-base...
-
Heterogeneous Graph Backdoor Attack
HGBA is a backdoor attack against heterogeneous graph neural networks that uses a single relation-based trigger edge to achieve high attack success with low budget and resistance to defenses.
-
Benchmarking Sheaf Neural Networks for Inductive Tasks
On 14 inductive graph benchmarks, sheaf neural networks underperform strong GNN baselines, and their performance is driven more by the surrounding architecture than by the sheaf diffusion mechanism.
-
MEGA-CL: A Molecular Foundation Model for Generalizable ADMET Prediction through Graph External Attention and Contrastive Learning
A contrastive GNN with external attention, pre-trained on ~100M molecules, improves multi-task ADMET prediction and shows prospective wet-lab agreement on three compounds.
-
Learnable quantum spectral filters for hybrid graph neural networks
A parameterized quantum Fourier circuit with graph-derived gate connections acts as a convolution plus pooling layer in a hybrid quantum-classical graph neural network, achieving benchmark accuracies comparable to som...
-
How do Probabilistic Graphical Models and Graph Neural Networks Look at Network Data?
SBM-style probabilistic models outperform graph neural networks on link prediction when node features are low-dimensional, noisy, or the graph is heterophilic.
-
Chordless Structure: A Pathway to Simple and Expressive GNNs
CSGNN encodes chordless cycles and paths as structural features and is claimed to strictly outperform k-hop GNNs in expressiveness with polynomial complexity.
-
Scalable inference of large-scale random kronecker graphs via tensor decomposition and Einstein summation
A tensor version of Kronecker-graph denoising and parameter inference is presented, but it reuses prior matrix results with a layer index and contains model-definition and proof gaps.
Discussion (0). Sign in to comment.