English
Related papers

Related papers: The Oversmoothing Fallacy: A Misguided Narrative i…

200 papers

Graph convolutional networks (GCNs) have shown promising results in processing graph data by extracting structure-aware features. This gave rise to extensive work in geometric deep learning, focusing on designing network architectures that…

Machine Learning · Computer Science 2022-01-20 Yimeng Min , Frederik Wenkel , Guy Wolf

In recent years, Graph Neural Networks (GNNs) have been popular in the graph classification task. Currently, shallow GNNs are more common due to the well-known over-smoothing problem facing deeper GNNs. However, they are sub-optimal without…

Machine Learning · Computer Science 2023-02-20 Lanning Wei , Zhiqiang He , Huan Zhao , Quanming Yao

Current GNN-oriented NAS methods focus on the search for different layer aggregate components with shallow and simple architectures, which are limited by the 'over-smooth' problem. To further explore the benefits from structural diversity…

Machine Learning · Computer Science 2021-09-22 Guosheng Feng , Chunnan Wang , Hongzhi Wang

Our study reveals new theoretical insights into over-smoothing and feature over-correlation in graph neural networks. Specifically, we demonstrate that with increased depth, node representations become dominated by a low-dimensional…

Machine Learning · Computer Science 2024-09-19 Andreas Roth , Thomas Liebig

While Graph Neural Networks (GNNs) are powerful models for learning representations on graphs, most state-of-the-art models do not have significant accuracy gain beyond two to three layers. Deep GNNs fundamentally need to address: 1).…

In this work, we generalize the ideas of Kaiming initialization to Graph Neural Networks (GNNs) and propose a new scheme (G-Init) that reduces oversmoothing, leading to very good results in node and graph classification tasks. GNNs are…

Machine Learning · Computer Science 2024-11-01 Dimitrios Kelesis , Dimitris Fotakis , Georgios Paliouras

Graph neural networks (GNNs) are able to leverage the structure of graph data by passing messages along the edges of the graph. While this allows GNNs to learn features depending on the graph structure, for certain graph topologies it leads…

Machine Learning · Computer Science 2023-02-17 Kedar Karhadkar , Pradeep Kr. Banerjee , Guido Montúfar

Message Passing Neural Networks (MPNNs) hold a key position in machine learning on graphs, but they struggle with unintended behaviors, such as over-smoothing and over-squashing, due to irregular data structures. The observation and…

Machine Learning · Computer Science 2025-08-26 Junhyun Lee , Veronika Thost , Bumsoo Kim , Jaewoo Kang , Tengfei Ma

Graph neural networks have shown significant success in the field of graph representation learning. Graph convolutions perform neighborhood aggregation and represent one of the most important graph operations. Nevertheless, one layer of…

Machine Learning · Computer Science 2020-07-21 Meng Liu , Hongyang Gao , Shuiwang Ji

There has been tremendous success in the field of graph neural networks (GNNs) as a result of the development of the message-passing (MP) layer, which updates the representation of a node by combining it with its neighbors to address…

Machine Learning · Computer Science 2022-02-11 Hyeokjin Kwon , Jong-Min Lee

Message Passing Neural Networks (MPNNs) is the building block of graph foundation models, but fundamentally suffer from oversmoothing and oversquashing. There has recently been a surge of interest in fixing both issues. Existing efforts…

Machine Learning · Computer Science 2025-10-21 Li Sun , Zhenhao Huang , Ming Zhang , Philip S. Yu

Graph Neural Networks (GNNs) set the state-of-the-art in representation learning for graph-structured data. They are used in many domains, from online social networks to complex molecules. Most GNNs leverage the message-passing paradigm and…

Machine Learning · Computer Science 2025-03-06 Tuğrul Hasan Karabulut , İnci M. Baytaş

Graph neural networks (GNNs) have become pivotal tools for processing graph-structured data, leveraging the message passing scheme as their core mechanism. However, traditional GNNs often grapple with issues such as instability,…

Spectral Theory · Mathematics 2026-05-20 Yuanhong Jiang , Dongmian Zou , Xiaoqun Zhang , Yu Guang Wang

Graph neural networks (GNNs) are one of the most popular research topics for deep learning. GNN methods typically have been designed on top of the graph signal processing theory. In particular, diffusion equations have been widely used for…

Machine Learning · Computer Science 2023-06-16 Jeongwhan Choi , Seoyoung Hong , Noseong Park , Sung-Bae Cho

Graph neural networks (GNNs) are designed to process data associated with graphs. They are finding an increasing range of applications; however, as with other modern machine learning techniques, their theoretical understanding is limited.…

Disordered Systems and Neural Networks · Physics 2026-02-23 O. Duranthon , L. Zdeborová

Graph Convolutional Network (GCN) with the powerful capacity to explore graph-structural data has gained noticeable success in recent years. Nonetheless, most of the existing GCN-based models suffer from the notorious over-smoothing issue,…

Machine Learning · Computer Science 2023-04-17 Zhaoliang Chen , Zhihao Wu , Zhenghong Lin , Shiping Wang , Claudia Plant , Wenzhong Guo

Graph neural networks (GNNs) have achieved significant success in various applications. Most GNNs learn the node features with information aggregation of its neighbors and feature transformation in each layer. However, the node features…

Machine Learning · Computer Science 2024-07-30 Dongzhuoran Zhou , Hui Yang , Bo Xiong , Yue Ma , Evgeny Kharlamov

Deep graph neural networks (GNNs) often suffer from oversmoothing, where node representations become overly homogeneous with increasing depth. While techniques like normalization, residual connections, and edge dropout have been proposed to…

Machine Learning · Computer Science 2025-05-30 Jiaqi Wang , Xinyi Wu , James Cheng , Yifei Wang

Graph Neural Networks (GNNs) have succeeded in various computer science applications, yet deep GNNs underperform their shallow counterparts despite deep learning's success in other domains. Over-smoothing and over-squashing are key…

Machine Learning · Computer Science 2023-08-14 Jhony H. Giraldo , Konstantinos Skianis , Thierry Bouwmans , Fragkiskos D. Malliaros

Deep Graph Neural Networks struggle with oversmoothing. This paper introduces a novel, physics-inspired GNN model designed to mitigate this issue. Our approach integrates with existing GNN architectures, introducing an entropy-aware message…

Machine Learning · Computer Science 2024-03-08 Philipp Nazari , Oliver Lemke , Davide Guidobene , Artiom Gesp