Lmst

#ontologicalengineering

In Weimar at the Workshop on Generative AI for Cultural Data and Text Data, I was presenting "Intuition vs Precision - LLMs for Ontological Engineering", in which we also show some experiments in how LLMs can support the ontology lifecycle.

presentation: https://zenodo.org/records/14041279

#llms #ontologies #ontologicalengineering #dh #digitalhumanities @nfdi4culture @NFDI4Memory #NFDIrocks @fizise @fiz_karlsruhe @enorouzi @sourisnumerique #semanticweb #knowledgegraphs

cover slide of my presentation "Intuition vs Precision - LLMs in Ontological Engineering", The development of precise and meaningful ontologies requires close collaboration between ontology experts and domain specialists. The necessary knowledge transfer between these groups is often time-consuming and complex. Large Language Models (LLMs) offer promising approaches to automate and accelerate this process. LLMs can support various phases of the ontology lifecycle, for example, in generating term suggestions, creating definitions, or generating and evaluating competence questions. The talk highlights various possible uses of LLMs in ontology development, discusses both opportunities and risks, and presents initial results.

Reflections on the Special Track on LLMs for Knowledge Engineering at #ESWC2024.
From my point of view a very valuable contribution to the conference, exchanging our experiences and research efforts as long as there are no established methodologies and evaluations in this field.

#llms #knowledgeengineering #knowledgegraphs #ontologydesign #ontologicalengineering ontologies

Reflections C How do we make a useful benchmark? * ©.g protecting the benchmark * Isthisoutof sample for the LLM? * LWMleaderboard * Useful outside of the community / reproducibility * Whatworks across / independent of LLMs * Where do we get data with some of the tasks? * Availability of data about the process of knowledge engineering * Speedand variability of the LLMs * Move towards rearchitecting KE workflows ; e Need more in-practice human evaluation * Whole workflow evaluation | * Return of methodoloe){ * Small selection of tasks -what are major tasks that are missing? * More outputs when using LLMs meaning more things to evaluate * The Surprising importance of syntax * Where/who are making the mistakes * Environmental impact * The Propogation of Potential bias from LLMs to the knowledge graphs . Documenting the learnings from these studies in astructured way * Studying the changing relationship between knoMedge engineer, LLM, KG . Potentially look at software engineering 2 £ v

FIrst #ESWC2024 Keynote by Elena Simperl on "Knowledge Engineering from people to machines and back" dives into the history of knowledge engineering and reminds us about the many different approaches ranging from people, over more people, gamification, crowdsourcing, automation, etc.

#knowledgeengineering #semanticweb #ontologies #ontologicalengineering #knowledgegraphs

Elena Simperl presenting her keynote at ESWC2024

We are glad to be part of the Innovation Platform MaterialDigital project, combining 24 consortia in their effort to digitalize #materialsscience engineering. Our task is to facilitate #ontology design for #mse experts who are not familiar with knowledge engineering.

video: https://www.youtube.com/watch?v=ilxazOFDHSk
MaterialDigital website: https://www.materialdigital.de/projects/

@fiz_karlsruhe @KIT_Karlsruhe @BAMResearch #materialdigital @joerg @heikef @enorouzi @fizise

#ontologies #ontologicalengineering #KnowledgeGraphs

video screenshot showing Prof. Dr. Peter Gumbsch from KIT, speaker of the Platform MaterialDigital project.

Zum nächstmöglichen Eintrittstermin suchen wir einen PhD/Junior Researcher oder PostDoc/Senior Researcher (w/m/x) im Information Service Engineering Forschungsteam (ISE).
👉 Mehr: https://lnkd.in/eXJ9cYeJ

Die Stelle ist an das BMBF Projekt „Innovationsplattform MaterialDigital II“ zur nachhaltigen #Digitalisierung der Material- und Werkstoffwissenschaften gebunden.
@fizise #job #stellenanzeige #karriere #KnowledgeGraphs #deeplearning #OntologicalEngineering

We are happy to welcome Sven Hertling on board as new FIZ ISE PostDoc team member. Sven comes from the research group of Heiko Paulheim at University of Mannheim and will continue his research together with us at @fiz_karlsruhe on #knowledgegraphs and #ontologicalEngineering #ontologyMatching #semanticweb #ai

GoogleScholar: https://scholar.google.com/citations?user=N7MADekAAAAJ&hl=en&oi=ao
DBLP: https://dblp.org/pid/129/9510.html

@tabea @sashabruns @sourisnumerique @MahsaVafaie @enorouzi @heikef @GenAsefa @epoz

Portrait picture of Sven, smiling. He's wearing a white shirt and a dark jacket.

Our new paper on the PMDcore ontology bridging semantic gaps in Materials Science and Engineering to enable data interoperability has been published in Materials&Design:
Bernd Bayerlein, Markus Schilling, Henk Birkholz, Matthias Jung, @joerg, Lutz Mädler, @lysander07 Core Ontology: Achieving semantic interoperability in materials science
paper link: https://www.sciencedirect.com/science/article/pii/S0264127523010195?via%3Dihub

@fiz_karlsruhe @fizise #ontologies @enorouzi #mse #knowledgegraphs #semanticweb #ontologicalengineering

A schematic graphics to visualize the contributions of the paper as highlighted in the abstract:

Knowledge representation in the Materials Science and Engineering (MSE) domain is a vast and multi-faceted challenge: Overlap, ambiguity, and inconsistency in terminology are common. Invariant (consistent) and variant (context-specific) knowledge are difficult to align cross-domain. Generic top-level semantic terminology often is too abstract, while MSE domain terminology often is too specific. In this paper, an approach how to maintain a comprehensive MSE-centric terminology composing a mid-level ontology–the Platform MaterialDigital Core Ontology (PMDco)–via MSE community-based curation procedures is presented. The illustrated findings show how the PMDco bridges semantic gaps between high-level, MSE-specific, and other science domain semantics. Additionally, it demonstrates how the PMDco lowers development and integration thresholds. Moreover, the research highlights how to fuel it with real-world data sources ranging from manually conducted experiments and simulations with continuously automated industrial applications.

The new TGDK website in online via our new publisher Dagstuhl. Transactions on Graph Data and Knowledge (TGDK) is a Diamond #OpenAccess journal that publishes research contributions relating to the use of graphs for data and knowledge management.

https://www.dagstuhl.de/en/publishing/series/details/TGDK

#knowledgegraphs #ontologies #knowledgegraph #semanticweb #graphdata #ontologicalengineering #llms #knowledgeextraction #knowledgemining ##tgdk @gdm @katjahose @ejimenez_ruiz @keet @catiapesquita @AxelPolleres @juan

Webpage of TGDK
Transactions on Graph Data and Knowledge (TGDK) is an Open Access journal that publishes original research articles and survey articles on graph-based abstractions for data and knowledge, and the techniques that such abstractions enable with respect to integration, querying, reasoning and learning. The scope of the journal thus intersects with areas such as Graph Algorithms, Graph Databases, Graph Representation Learning, Knowledge Graphs, Knowledge Representation, Linked Data and the Semantic Web. Also in-scope for the journal is research investigating graph-based abstractions of data and knowledge in the context of Data Integration, Data Science, Information Extraction, Information Retrieval, Machine Learning, Natural Language Processing, and the Web.

How to deal with different ontologies on the same subject? What can we do to get them aligned? Is it possible to learn an ontology from text? These questions, we are going to address in this section of our #kg2023 lecture with the title "Ontological Engineering"
#OpenHPI lecture: https://open.hpi.de/courses/knowledgegraphs2023/items/VseJ7SbPrJS5t8qgHuv57
youtube lecture: https://www.youtube.com/watch?v=w0JmROLn_OA&list=PLNXdQl4kBgzubTOfY5cbtxZCgg9UTe-uF&index=52
slides: https://zenodo.org/records/10135498
@fiz_karlsruhe @fizise @tabea @sashabruns @MahsaVafaie #knowledgegraphs #ontologies #ontologicalengineering #semanticweb

Cover Slide of the OpenHPI lecture Knowledge Graphs - Foundations and Applications, Week 5: Ontological Engineering for smarter Knowledge Graphs, 5.4 Ontological Engineering. The background shows a picture generated by artbot, prompt: “On this scifi movie poster we see the vibrant construction site of a gigantic space ship in the vast deserts of planet Mars exposing many small details.”, created via ArtBot, Deliberate, 2023, [CC-BY-4.0], https://tinybots.net/artbot

Today, we had a super interesting ISE research seminar presentation from our colleague @sashabruns on variants of temporal modeling for RDF, as e.g. RDF reification, RDF singleton properties, named graphs, RDF*, and more. How do you evaluate the efficiency of the different modelings? Guidelines and best practices are currently developed.
@fizise @tabea @enorouzi @heikef @joerg #RDF #temporalLogic #semanticweb #knowledgegraphs #ontologies #ontologicalengineering

Cover slide of Sasha Bruns ISE research seminar presentation "Oh Honey, or How Naive is your Temporal Modeling?"

Client Info

Server: https://mastodon.social

Version: 2025.07

Repository: https://github.com/cyevgeniy/lmst