Search Results

Knowledge engineering is a discipline that focuses on the creation and maintenance of processes that generate and apply knowledge. Traditionally, knowledge engineering approaches have focused on knowledge expressed in formal languages. The emergence of large language models and their capabilities to effectively work with natural language, in its broadest sense, raises questions about the foundations and practice of knowledge engineering. Here, we outline the potential role of LLMs in knowledge engineering, identifying two central directions: 1) creating hybrid neuro-symbolic knowledge systems; and 2) enabling knowledge engineering in natural language. Additionally, we formulate key open research questions to tackle these directions.

This report documents the programme and outcomes of Dagstuhl Seminar 22372 "Knowledge Graphs and their Role in the Knowledge Engineering of the 21st Century" held in September 2022. The seminar aimed to gain a better understanding of the way knowledge graphs are created, maintained, and used today, and identify research challenges throughout the knowledge engineering life cycle, including tasks such as modelling, representation, reasoning, and evolution. The participants identified directions of research to answer these challenges, which will form the basis for new methodologies, methods, and tools, applicable to varied AI systems in which knowledge graphs are used, for instance, in natural language processing, or in information retrieval. The seminar brought together a snapshot of the knowledge engineering and adjacent communities, including leading experts, academics, practitioners, and rising stars in those fields. It fulfilled its aims - the participants took inventory of existing and emerging solutions, discussed open problems and practical challenges, and identified ample opportunities for novel research, technology transfer, and inter-disciplinary collaborations. Among the topics of discussion were: designing engineering methodologies for knowledge graphs, integrating large language models and structured data into knowledge engineering pipelines, neural methods for knowledge engineering, responsible use of AI in knowledge graph construction, other forms of knowledge representations, and generating user and developer buy-in. Besides a range of joint publications, hackathons, and project proposals, the participants suggested joint activities with other scientific communities, in particular those working on large language models, generative AI, FAccT (fairness, accountability, transparency), and human-AI interaction. The discussions were captured in visual summaries thanks to Catherine Allan - you can find more about her work at https://www.catherineallan.co.uk/. The summaries are arrayed throughout this report. Lastly, knowledge about the seminar is captured in Wikidata at https://www.wikidata.org/wiki/Q113961931