REGE: A Method for Incorporating Uncertainty in Graph Embeddings

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shafi, Zohair, Savcisens, Germans, Eliassi-Rad, Tina
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929619837386752
author Shafi, Zohair
Savcisens, Germans
Eliassi-Rad, Tina
author_facet Shafi, Zohair
Savcisens, Germans
Eliassi-Rad, Tina
contents Machine learning models for graphs in real-world applications are prone to two primary types of uncertainty: (1) those that arise from incomplete and noisy data and (2) those that arise from uncertainty of the model in its output. These sources of uncertainty are not mutually exclusive. Additionally, models are susceptible to targeted adversarial attacks, which exacerbate both of these uncertainties. In this work, we introduce Radius Enhanced Graph Embeddings (REGE), an approach that measures and incorporates uncertainty in data to produce graph embeddings with radius values that represent the uncertainty of the model's output. REGE employs curriculum learning to incorporate data uncertainty and conformal learning to address the uncertainty in the model's output. In our experiments, we show that REGE's graph embeddings perform better under adversarial attacks by an average of 1.5% (accuracy) against state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2412_05735
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle REGE: A Method for Incorporating Uncertainty in Graph Embeddings
Shafi, Zohair
Savcisens, Germans
Eliassi-Rad, Tina
Machine Learning
Machine learning models for graphs in real-world applications are prone to two primary types of uncertainty: (1) those that arise from incomplete and noisy data and (2) those that arise from uncertainty of the model in its output. These sources of uncertainty are not mutually exclusive. Additionally, models are susceptible to targeted adversarial attacks, which exacerbate both of these uncertainties. In this work, we introduce Radius Enhanced Graph Embeddings (REGE), an approach that measures and incorporates uncertainty in data to produce graph embeddings with radius values that represent the uncertainty of the model's output. REGE employs curriculum learning to incorporate data uncertainty and conformal learning to address the uncertainty in the model's output. In our experiments, we show that REGE's graph embeddings perform better under adversarial attacks by an average of 1.5% (accuracy) against state-of-the-art methods.
title REGE: A Method for Incorporating Uncertainty in Graph Embeddings
topic Machine Learning
url https://arxiv.org/abs/2412.05735