We are excited to announce that four papers authored by SemSys members have been accepted at ESWC 2024! Three papers investigate the use of Large Language Models in knowledge engineering contexts, with a strong focus on the validation and evaluation of semantic artefacts, while another one demonstrates an approach for cataloguing semantic web tools.

Special Track on LLMs for Knowledge Engineering:

  1. Validating Semantic Artefacts With Large Language Models 

Authors: Tufek, N., Thuluva, A.S., Bandyopadhyay, T., Just, V.P., Sabou, M.,  Ekaputra, F. J., and Hanbury, A.  

This work is led by colleagues from Siemens and in collaboration with colleagues from TU Wien. It proposes a method to use LLMs for translating natural language questions (such as competency questions) into a machine-processable SPARQL representation. These can be used to automatically validate semantic artefacts such as ontologies, knowledge graphs or semantic models represented in industrial standards such as OPC-UA. 

  1. OntoChat: a Framework for Conversational Ontology Engineering using Language Models.

Authors: Zhang, B., Carriero, V.A., Schreiberhuber, K., Tsaneva, S., Gonz√°lez, L.S., Kim, J. and de Berardinis, J. 

This work is a continuation of a group project initiated during the Knowledge Prompting Hackathon which took place last August in London. The paper introduces OntoChat, an LLM-based framework supporting various stakeholders across different ontology engineering stages (from user story generation to ontology testing). The framework is available online (OntoChat) and the implementation is released on Github.     

Workshop on Data Quality meets Machine Learning and Knowledge Graphs (DQMLKG)

  1. LLM-driven Ontology Evaluation: Verifying Ontology Restrictions with ChatGPT 

Authors: Tsaneva S., Vasic S., Sabou M. 

In the paper we replicate a previously conducted human-in-the-loop experiment where semi-experts were tasked with verifying the correct usage of ontology restrictions. In the replication, elements of the original human intelligence tasks are used to design ChatGPT prompts, resulting in verification accuracy of up to 96.67%. 

ESWC Poster & Demo Track:  

  1. Semantic Tool Hub: Towards A Sustainable Community-Driven Documentation of Semantic Web Tools
    Authors: Reiz A., Ekaputra, FJ. Mihindukulasooriya, N. 

This demo paper proposed an extended workflow and toolkit developed around Wikidata and GitHub to support knowledge engineers and SW tool developers in finding the right tools for their use and documenting SW tools that they have developed.

We are looking forward to attending ESWC 2024!