GSINA: Improving Subgraph Extraction for Graph Invariant Learning via Graph Sinkhorn Attention

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Yan, Junchi, Ding, Fangyu, Sun, Jiawei, Hu, Zhaoping, Zhou, Yunyi, Zhu, Lei
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911387881570304
author Yan, Junchi
Ding, Fangyu
Sun, Jiawei
Hu, Zhaoping
Zhou, Yunyi
Zhu, Lei
author_facet Yan, Junchi
Ding, Fangyu
Sun, Jiawei
Hu, Zhaoping
Zhou, Yunyi
Zhu, Lei
contents Graph invariant learning (GIL) seeks invariant relations between graphs and labels under distribution shifts. Recent works try to extract an invariant subgraph to improve out-of-distribution (OOD) generalization, yet existing approaches either lack explicit control over compactness or rely on hard top-$k$ selection that shrinks the solution space and is only partially differentiable. In this paper, we provide an in-depth analysis of the drawbacks of some existing works and propose a few general principles for invariant subgraph extraction: 1) separability, as encouraged by our sparsity-driven mechanism, to filter out the irrelevant common features; 2) softness, for a broader solution space; and 3) differentiability, for a soundly end-to-end optimization pipeline. Specifically, building on optimal transport, we propose Graph Sinkhorn Attention (GSINA), a fully differentiable, cardinality-constrained attention mechanism that assigns sparse-yet-soft edge weights via Sinkhorn iterations and induces node attention. GSINA provides explicit controls for separability and softness, and uses a Gumbel reparameterization to stabilize training. It convergence behavior is also theoretically studied. Extensive empirical experimental results on both synthetic and real-world
format Preprint
id arxiv_https___arxiv_org_abs_2402_07191
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle GSINA: Improving Subgraph Extraction for Graph Invariant Learning via Graph Sinkhorn Attention
Yan, Junchi
Ding, Fangyu
Sun, Jiawei
Hu, Zhaoping
Zhou, Yunyi
Zhu, Lei
Machine Learning
Artificial Intelligence
Graph invariant learning (GIL) seeks invariant relations between graphs and labels under distribution shifts. Recent works try to extract an invariant subgraph to improve out-of-distribution (OOD) generalization, yet existing approaches either lack explicit control over compactness or rely on hard top-$k$ selection that shrinks the solution space and is only partially differentiable. In this paper, we provide an in-depth analysis of the drawbacks of some existing works and propose a few general principles for invariant subgraph extraction: 1) separability, as encouraged by our sparsity-driven mechanism, to filter out the irrelevant common features; 2) softness, for a broader solution space; and 3) differentiability, for a soundly end-to-end optimization pipeline. Specifically, building on optimal transport, we propose Graph Sinkhorn Attention (GSINA), a fully differentiable, cardinality-constrained attention mechanism that assigns sparse-yet-soft edge weights via Sinkhorn iterations and induces node attention. GSINA provides explicit controls for separability and softness, and uses a Gumbel reparameterization to stabilize training. It convergence behavior is also theoretically studied. Extensive empirical experimental results on both synthetic and real-world
title GSINA: Improving Subgraph Extraction for Graph Invariant Learning via Graph Sinkhorn Attention
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2402.07191