Language Model Analysis for Ontology Subsumption Inference
This work addresses the gap in applying language models to sophisticated conceptual knowledge bases, which is incremental as it extends existing probing methods to a new domain.
The paper tackles the problem of whether pre-trained language models can serve as knowledge bases for complex logic-based ontologies, specifically OWL ontologies, and finds that while they encode less knowledge for subsumption inference than for natural language inference, they improve significantly with few-shot learning.
Investigating whether pre-trained language models (LMs) can function as knowledge bases (KBs) has raised wide research interests recently. However, existing works focus on simple, triple-based, relational KBs, but omit more sophisticated, logic-based, conceptualised KBs such as OWL ontologies. To investigate an LM's knowledge of ontologies, we propose OntoLAMA, a set of inference-based probing tasks and datasets from ontology subsumption axioms involving both atomic and complex concepts. We conduct extensive experiments on ontologies of different domains and scales, and our results demonstrate that LMs encode relatively less background knowledge of Subsumption Inference (SI) than traditional Natural Language Inference (NLI) but can improve on SI significantly when a small number of samples are given. We will open-source our code and datasets.