A generalization of vit/mlp-mixer to graphs

X He, B Hooi, T Laurent, A Perold… - International …, 2023 - proceedings.mlr.press
Abstract Graph Neural Networks (GNNs) have shown great potential in the field of graph
representation learning. Standard GNNs define a local message-passing mechanism which …

How powerful are k-hop message passing graph neural networks

J Feng, Y Chen, F Li, A Sarkar… - Advances in Neural …, 2022 - proceedings.neurips.cc
The most popular design paradigm for Graph Neural Networks (GNNs) is 1-hop message
passing---aggregating information from 1-hop neighbors repeatedly. However, the …

Rethinking the expressive power of gnns via graph biconnectivity

B Zhang, S Luo, L Wang, D He - arXiv preprint arXiv:2301.09505, 2023 - arxiv.org
Designing expressive Graph Neural Networks (GNNs) is a central topic in learning graph-
structured data. While numerous approaches have been proposed to improve GNNs in …

Weisfeiler and leman go machine learning: The story so far

C Morris, Y Lipman, H Maron, B Rieck… - The Journal of Machine …, 2023 - dl.acm.org
In recent years, algorithms and neural architectures based on the Weisfeiler-Leman
algorithm, a well-known heuristic for the graph isomorphism problem, have emerged as a …

A complete expressiveness hierarchy for subgraph gnns via subgraph weisfeiler-lehman tests

B Zhang, G Feng, Y Du, D He… - … Conference on Machine …, 2023 - proceedings.mlr.press
Recently, subgraph GNNs have emerged as an important direction for developing
expressive graph neural networks (GNNs). While numerous architectures have been …

Facilitating graph neural networks with random walk on simplicial complexes

C Zhou, X Wang, M Zhang - Advances in Neural …, 2024 - proceedings.neurips.cc
Node-level random walk has been widely used to improve Graph Neural Networks.
However, there is limited attention to random walk on edge and, more generally, on $ k …

Ordered subgraph aggregation networks

C Qian, G Rattan, F Geerts… - Advances in Neural …, 2022 - proceedings.neurips.cc
Numerous subgraph-enhanced graph neural networks (GNNs) have emerged recently,
provably boosting the expressive power of standard (message-passing) GNNs. However …

Equivariant polynomials for graph neural networks

O Puny, D Lim, B Kiani, H Maron… - … on Machine Learning, 2023 - proceedings.mlr.press
Abstract Graph Neural Networks (GNN) are inherently limited in their expressive power.
Recent seminal works (Xu et al., 2019; Morris et al., 2019b) introduced the Weisfeiler …

Universal prompt tuning for graph neural networks

T Fang, Y Zhang, Y Yang, C Wang… - Advances in Neural …, 2024 - proceedings.neurips.cc
In recent years, prompt tuning has sparked a research surge in adapting pre-trained models.
Unlike the unified pre-training strategy employed in the language field, the graph field …

Rethinking tokenizer and decoder in masked graph modeling for molecules

Z Liu, Y Shi, A Zhang, E Zhang… - Advances in …, 2024 - proceedings.neurips.cc
Masked graph modeling excels in the self-supervised representation learning of molecular
graphs. Scrutinizing previous studies, we can reveal a common scheme consisting of three …