← OntoLM tutorial (ISWC 2026)

Hands-on demos

Two short narrated walkthroughs of Part II. Each one runs a Colab notebook step by step on a small family ontology. Open the notebook to try it yourself.

OnT: train, then score

Fine-tune a language model on an OWL ontology with OntologyTransformer.fit(), then score subsumption axioms C ⊑ D with score_hierarchy().

  • Setup
  • Ontology
  • 1 · Fine-tune
  • 2 · Score
  • Before vs after
  • Parameters
Open in Colab Notebook on GitHub 3:48 · runs on the free Colab CPU

BERTSubs with DeepOnto

Call BERTSubs directly through DeepOnto: fine-tune BERT on the ontology's own subclass axioms, then read the probability of new ones.

  • Setup
  • Ontology
  • 1 · Configure
  • 2 · Run
  • 3 · Read
  • Before vs after
  • Options
Open in Colab Notebook on GitHub 3:25 · use a Colab GPU runtime