Generating text from ontologies in multiple languages
This page contains information and materials for the half-day tutorial on natural language generation with ontologies, by Maria Keet, with the University of Cape Town, and Zola Mahlaza, with the University of Pretoria, South Africa, at the Joint Ontology Workshops 2022. JOWO'22, which is an umbrella event with a collection of co-located workshops and tutorials on various aspect of ontologies, will take place at Jönköping University in Jönköping, Sweden, from 15-19 August 2022.
Tutorial venue and time
Friday 19 August 2022, 11:00-12:30 (part I, foundations) and 14:00-15:30 (part II, advanced topics). Both parts will be held in "room 2".
The principal aim is to introduce the participant to controlled natural languages (CNLs) and natural language generation (NLG). This will include also sub-topics such as key design decisions when verbalising ontologies and the knock-on effects of how the knowledge is represented in ontologies, multiple natural languages, architectures and examples of systems, and research challenges.
There will also be a hands-on part, where the participants will create templates for different types of axioms and for different purposes, apply them to an ontology of choice and one from the pre-selected ones, and reflect on evaluation of ontology verbalisation to obtain good quality sentences.
The intended audience are mainly postgraduate students and interested academics who would like to gain an appreciation of ontology verbalisation and NLG beyond repeating the limited low-hanging fruit.
Background/prerequisites: participants will gain more from the tutorial if they have a basic foundation in ontologies and OWL.
The course material consists of:
- Lecture slides: in pdf;
- Reading material for the tutorial, before or after: see below
- Materials for the hands-on session:
Supplementary materialAnnotated other resources you may find of interest:
- Chapter 9 of the introduction to ontology engineering textbook, on ontologies and natural language
- A list of ontology verbalisers with some key features, being the input they take, realisation method, natural language they generate, and to what extent they use a grammar as well
- Natural language generation for the Semantic Web, a survey: N. Bouayad-Agha, G. Casamayor, and L. Wanner. Natural language generation in the context of the semantic web. Semantic Web Journal, 2014, 5(6): 493-513.
- Controlled natural languages for ontology verbalisation, a survey: Hazem Safwat and Brian Davis. CNLs for the semantic web: a state of the art. Language Resources & Evaluation, 2017, 51(1):191-220.
- Natural language generation survey (i.e., not tailored to ontologies): Albert Gatt and Emiel Krahmer. Survey of the state of the art in natural language generation: core tasks, applications and evaluation. J. Artif. Int. Res., 2018, 61(1): 65–170.
- Controlled natural languages survey (i.e., not tailored to ontologies): Tobias Kuhn. A Survey and Classification of Controlled Natural Languages. Computational Linguistics, 40(1), 2014.
- An old paper on NLG, but useful as a reference framework for the tasks in the pipeline: E. Reiter and R. Dale. Building applied natural language generation systems. Natural Language Engineering, 1997, 3:57-87.
- A model in OWL, called ToCT, for specifying CNL templates
- Abstract Wikipedia, which tries to do the NLG at scale from a knowledge graph together with WikiFunctions and WikiData
- Early advances on the multilingual semantic web: Paul Buitelaar and Philipp Cimiano, editors. Towards the Multilingual Semantic Web: Principles, Methods and Applications. Springer, 2014.
- The ontolex-lemon community specification for multilingual ontologies and linked data
- A tool for localisation of the Protégé ODE into another language, illustrated with Spanish and Afrikaans
- A recent sample usage of ontology verbalisation: to generate answerable questions from ontologies for educational exercises
- A not-so-recent sample usage of ontology verbalisation in the digital humanities: to generate museum artefact descriptions from ontologies in Greek
- (to be extended)
Maria Keet: mkeet -at- cs.uct.ac.za
Zola Mahlaza: z.mahlaza -at- up.ac.za