When LLMs Encounter Open-world Graph Learning: A Fresh View on Unlabeled Data Uncertainty

ICLR 2026 Conference Submission15815 Authors

Keywords: Open-world Graph Learning, Unknown-class Rejection, Open-world Graph Assistant

TL;DR: We propose OGA, an LLM-enhanced open-world graph learning pipeline that unifies unknown-class rejection and annotation, effectively addressing unlabeled data uncertainty in text-attributed graphs.

Abstract: Recently, large language models (LLMs) have driven a systematic shift in the graph ML community through the adoption of text-attributed graphs (TAGs). Although a variety of frameworks have been developed, most fail to properly address the challenge of data uncertainty in open-world environments, which is vital for real-world deployment. A representative source of such uncertainty is the limited availability of labels in large-scale datasets due to high annotation costs, where unlabeled nodes may either belong to known classes or represent novel, unknown classes. While node-level out-of-distribution detection and conventional open-world graph learning attempt to tackle this problem, two core limitations remain: Insufficient methods — TAGs integrate textual and structural information, yet existing approaches typically optimize semantics or topology in isolation for unknown-class rejection, limiting their effectiveness; \ding{173} Incomplete pipelines — handling unknown-class nodes is essential for model re-updates and long-term deployment, but most studies conduct only idealized analyses, such as assuming a predefined number of unknown classes, which restricts practical utility. To overcome these issues, we introduce the Open-world Graph Assistant (OGA), an LLM-based framework. OGA first performs unknown-class rejection via adaptive label traceability (ALT), harmoniously combining semantic and topological cues, and then applies the graph label annotator (GLA) for unknown-class annotation, allowing unlabeled nodes to contribute to model training. In essence, OGA offers a new pipeline that fully automates the handling of unlabeled nodes in open-world environments, and we establish a systematic benchmark covering four key aspects to validate its effectiveness and practicality through extensive experiments.

Primary Area: foundation or frontier models, including LLMs

Submission Number: 15815