Recognizing Predictive Substructures With Subgraph Information Bottleneck

Yu, Junchi; Xu, Tingyang; Rong, Yu; Bian, Yatao; Huang, Junzhou; He, Ran

doi:10.1109/tpami.2021.3112205

Cited by 36 publications

(24 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The GIB method enjoys several strength with the information-theoretical objective. First, the IB-Subgraph is proved to be noise-invariant [149]. Thus, it improves the robustness of the GNN model on graph classification task by dropping the redundancy and noisy information.…”

Section: Self-explainable Methodsmentioning

confidence: 99%

A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protection

Wu¹,

Li²,

Yu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Deep graph learning has achieved remarkable progresses in both business and scientific areas ranging from finance and e-commerce, to drug and advanced material discovery. Despite these progresses, how to ensure various deep graph learning algorithms behave in a socially responsible manner and meet regulatory compliance requirements becomes an emerging problem, especially in risk-sensitive domains. Trustworthy graph learning (TwGL) aims to solve the above problems from a technical viewpoint. In contrast to conventional graph learning research which mainly cares about model performance, TwGL considers various reliability and safety aspects of the graph

show abstract

Section: Self-explainable Methodsmentioning

confidence: 99%

A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protection

Wu¹,

Li²,

Yu³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…The IB framework has received significant attention in machine learning and deep learning (Alemi et al 2016;Saxe et al 2019). As for irregular graph data, there are some recent works (Wu et al 2020;Yu et al 2020;Yang et al 2021;Yu et al 2021) introducing the IB principle to graph learning. GIB (Wu et al 2020) extends the general IB to graph data with regularization of the structure and feature information for robust node representations.…”

Section: Information Bottleneckmentioning

confidence: 99%

Graph Structure Learning with Variational Information Bottleneck

Sun

Peng

et al. 2022

AAAI

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) have shown promising results on a broad spectrum of applications. Most empirical studies of GNNs directly take the observed graph as input, assuming the observed structure perfectly depicts the accurate and complete relations between nodes. However, graphs in the real-world are inevitably noisy or incomplete, which could even exacerbate the quality of graph representations. In this work, we propose a novel Variational Information Bottleneck guided Graph Structure Learning framework, namely VIB-GSL, in the perspective of information theory. VIB-GSL is the first attempt to advance the Information Bottleneck (IB) principle for graph structure learning, providing a more elegant and universal framework for mining underlying task-relevant relations. VIB-GSL learns an informative and compressive graph structure to distill the actionable information for specific downstream tasks. VIB-GSL deduces a variational approximation for irregular graph data to form a tractable IB objective function, which facilitates training stability. Extensive experimental results demonstrate that the superior effectiveness and robustness of the proposed VIB-GSL.

show abstract

“…The IB framework has received significant attention in machine learning and deep learning (Alemi et al 2016;Saxe et al 2019). As for irregular graph data, there are some recent works (Wu et al 2020;Yang et al 2021;Yu et al 2021) introducing the IB principle to graph learning. GIB (Wu et al 2020) extends the general IB to graph data with regularization of the structure and feature information for robust node representations.…”

Section: Information Bottleneckmentioning

confidence: 99%

“…GIB (Wu et al 2020) extends the general IB to graph data with regularization of the structure and feature information for robust node representations. SIB (Yu et al , 2021 was proposed for the subgraph recognition problem. HGIB (Yang et al 2021) was proposed to implement the consensus hypothesis of heterogeneous information networks in an unsupervised manner.…”

Section: Information Bottleneckmentioning

confidence: 99%

Graph Structure Learning with Variational Information Bottleneck

Sun¹,

Li²,

Peng³

et al. 2021

Preprint

View full text Add to dashboard Cite

Graph Neural Networks (GNNs) have shown promising results on a broad spectrum of applications. Most empirical studies of GNNs directly take the observed graph as input, assuming the observed structure perfectly depicts the accurate and complete relations between nodes. However, graphs in the real-world are inevitably noisy or incomplete, which could even exacerbate the quality of graph representations. In this work, we propose a novel Variational Information Bottleneck guided Graph Structure Learning framework, namely VIB-GSL, in the perspective of information theory. VIB-GSL advances the Information Bottleneck (IB) principle for graph structure learning, providing a more elegant and universal framework for mining underlying task-relevant relations. VIB-GSL learns an informative and compressive graph structure to distill the actionable information for specific downstream tasks. VIB-GSL deduces a variational approximation for irregular graph data to form a tractable IB objective function, which facilitates training stability. Extensive experimental results demonstrate that the superior effectiveness and robustness of VIB-GSL.

show abstract

Recognizing Predictive Substructures With Subgraph Information Bottleneck

Cited by 36 publications

References 26 publications

A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protection

A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protection

Graph Structure Learning with Variational Information Bottleneck

Graph Structure Learning with Variational Information Bottleneck

Contact Info

Product

Resources

About