GRAPHLCP: Structure-Aware Localized Conformal Prediction on Graphs

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Baghershahi, Peyman, Wang, Fangxin, Mandal, Debmalya, Medya, Sourav
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866913103889825792
author Baghershahi, Peyman
Wang, Fangxin
Mandal, Debmalya
Medya, Sourav
author_facet Baghershahi, Peyman
Wang, Fangxin
Mandal, Debmalya
Medya, Sourav
contents Conformal prediction (CP) provides a distribution-free approach to uncertainty quantification with finite-sample guarantees. However, applying CP to graph neural networks (GNNs) remains challenging as the combinatorial nature of graphs often leads to insufficiently certain predictions and indiscriminative embeddings. Existing methods primarily rely on embedding-space proximity for localization, which can be unreliable for graphs and yield inefficient prediction sets. We propose GRAPHLCP, a proximity-based localized CP framework that explicitly incorporates graph topology and inter-node dependencies into localization and weighting. Our approach introduces a feature-aware densification step to mitigate locality bias in sparse graphs, followed by a Personalized PageRank-based kernel computation to model structural proximity. This enables topology-dependent anchor sampling and calibration weighting that captures both local and long-range dependencies. Extensive experiments on several regression and classification datasets demonstrate that GRAPHLCP guarantees marginal coverage with finite samples while efficiently attaining favorable test conditional coverage across various conditioning scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2605_08074
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle GRAPHLCP: Structure-Aware Localized Conformal Prediction on Graphs
Baghershahi, Peyman
Wang, Fangxin
Mandal, Debmalya
Medya, Sourav
Machine Learning
Conformal prediction (CP) provides a distribution-free approach to uncertainty quantification with finite-sample guarantees. However, applying CP to graph neural networks (GNNs) remains challenging as the combinatorial nature of graphs often leads to insufficiently certain predictions and indiscriminative embeddings. Existing methods primarily rely on embedding-space proximity for localization, which can be unreliable for graphs and yield inefficient prediction sets. We propose GRAPHLCP, a proximity-based localized CP framework that explicitly incorporates graph topology and inter-node dependencies into localization and weighting. Our approach introduces a feature-aware densification step to mitigate locality bias in sparse graphs, followed by a Personalized PageRank-based kernel computation to model structural proximity. This enables topology-dependent anchor sampling and calibration weighting that captures both local and long-range dependencies. Extensive experiments on several regression and classification datasets demonstrate that GRAPHLCP guarantees marginal coverage with finite samples while efficiently attaining favorable test conditional coverage across various conditioning scenarios.
title GRAPHLCP: Structure-Aware Localized Conformal Prediction on Graphs
topic Machine Learning
url https://arxiv.org/abs/2605.08074