REVIEW 8 cited by
Is Heterophily A Real Nightmare For Graph Neural Networks To Do Node 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
Graph Neural Networks (GNNs) extend basic Neural Networks (NNs) by using the graph structures based on the relational inductive bias (homophily assumption). Though GNNs are believed to outperform NNs in real-world tasks, performance advantages of GNNs over graph-agnostic NNs seem not generally satisfactory. Heterophily has been considered as a main cause and numerous works have been put forward to address it. In this paper, we first show that not all cases of heterophily are harmful for GNNs with aggregation operation. Then, we propose new metrics based on a similarity matrix which considers the influence of both graph structure and input features on GNNs. The metrics demonstrate advantages over the commonly used homophily metrics by tests on synthetic graphs. From the metrics and the observations, we find some cases of harmful heterophily can be addressed by diversification operation. With this fact and knowledge of filterbanks, we propose the Adaptive Channel Mixing (ACM) framework to adaptively exploit aggregation, diversification and identity channels in each GNN layer to address harmful heterophily. We validate the ACM-augmented baselines with 10 real-world node classification tasks. They consistently achieve significant performance gain and exceed the state-of-the-art GNNs on most of the tasks without incurring significant computational burden.
Forward citations
Cited by 8 Pith papers
-
Same Graph Cross-Task Transfer in GNNs: Protocols and Predictors
Under a fixed leakage-free protocol, NC→LP transfer reliably helps on homophilic graphs while LP→NC helps mainly when LP is easy and NC is unsaturated; homophily and CoTask Score guide mechanism choice.
-
Dynamic Triangulation-Based Graph Rewiring for Graph Neural Networks
A learned triangle-selection module rewires graphs for GNNs, improving node classification over prior rewiring methods on 9 of 10 benchmarks.
-
ReFill: Reinforcement Learning for Fill-In Minimization
A reinforcement learning agent that picks variable elimination orders using graph neural networks slightly reduces fill-in versus minimum degree and minimum fill-in heuristics on small test graphs.
-
Partitioning Message Passing for Graph Fraud Detection
PMP partitions message passing by neighbor label and generates node-specific weights, reporting strong fraud detection results but with an invalid spectral proof.
-
Adapting to Heterophilic Graph Data with Structure-Guided Neighbor Discovery
Structure-Guided GNN combines the original graph with k-NN graphs built from role and global structural attributes and learns per-graph weights, achieving top results on 10 of 11 node-classification datasets.
-
THeGCN: Temporal Heterophilic Graph Convolutional Network
THeGCN uses learned low/high-pass attention over sampled temporal events to improve semi-supervised node classification on event-based continuous graphs with both edge and temporal heterophily.
-
IMPA-HGAE:Intra-Meta-Path Augmented Heterogeneous Graph Autoencoder
IMPA-HGAE masks and reconstructs features and meta-paths in heterogeneous graphs and propagates intermediate meta-path node information, yielding competitive but not uniformly best node-classification results across f...
-
Revisiting the Necessity of Graph Learning and Common Graph Benchmarks
Tuned feature-only MLPs nearly match graph neural networks on five common graph benchmarks, suggesting those benchmarks measure feature quality more than graph learning.
Discussion (0). Continue with ORCID to comment.