Graph Sparsification for Enhanced Conformal Prediction in Graph Neural Networks

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: He, Yuntian, Maneriker, Pranav, Srinivasan, Anutam, Vadlamani, Aditya T., Parthasarathy, Srinivasan
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910674443042816
author He, Yuntian
Maneriker, Pranav
Srinivasan, Anutam
Vadlamani, Aditya T.
Parthasarathy, Srinivasan
author_facet He, Yuntian
Maneriker, Pranav
Srinivasan, Anutam
Vadlamani, Aditya T.
Parthasarathy, Srinivasan
contents Conformal Prediction is a robust framework that ensures reliable coverage across machine learning tasks. Although recent studies have applied conformal prediction to graph neural networks, they have largely emphasized post-hoc prediction set generation. Improving conformal prediction during the training stage remains unaddressed. In this work, we tackle this challenge from a denoising perspective by introducing SparGCP, which incorporates graph sparsification and a conformal prediction-specific objective into GNN training. SparGCP employs a parameterized graph sparsification module to filter out task-irrelevant edges, thereby improving conformal prediction efficiency. Extensive experiments on real-world graph datasets demonstrate that SparGCP outperforms existing methods, reducing prediction set sizes by an average of 32\% and scaling seamlessly to large networks on commodity GPUs.
format Preprint
id arxiv_https___arxiv_org_abs_2410_21618
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Graph Sparsification for Enhanced Conformal Prediction in Graph Neural Networks
He, Yuntian
Maneriker, Pranav
Srinivasan, Anutam
Vadlamani, Aditya T.
Parthasarathy, Srinivasan
Machine Learning
Conformal Prediction is a robust framework that ensures reliable coverage across machine learning tasks. Although recent studies have applied conformal prediction to graph neural networks, they have largely emphasized post-hoc prediction set generation. Improving conformal prediction during the training stage remains unaddressed. In this work, we tackle this challenge from a denoising perspective by introducing SparGCP, which incorporates graph sparsification and a conformal prediction-specific objective into GNN training. SparGCP employs a parameterized graph sparsification module to filter out task-irrelevant edges, thereby improving conformal prediction efficiency. Extensive experiments on real-world graph datasets demonstrate that SparGCP outperforms existing methods, reducing prediction set sizes by an average of 32\% and scaling seamlessly to large networks on commodity GPUs.
title Graph Sparsification for Enhanced Conformal Prediction in Graph Neural Networks
topic Machine Learning
url https://arxiv.org/abs/2410.21618