Talk by Paul Groth on "Data Taste: Foregrounding Choices in Knowledge Graph Construction"

Dear all,

Happy New Year!

Within the GOBLIN Cost Action, we are organizing monthly Lunch Lecture Series [1]. The first talk will be given by Paul Groth (Professor in the University of Amsterdam, the Netherlands) on January 23, 2026 at 13:00. 

If you are interested in attending please register on [2] by January 20, 2026. Visit [1] for staying updated with the upcoming talks.

Title: Data Taste: Foregrounding Choices in Knowledge Graph Construction 

Abstract: Large Language Models (LLMs) have provided a powerful new capabilities for knowledge graph construction. In particular, they have dramatically decreased the cost of knowledge acquisition (e.g. information extraction and knowledge probing from LLMs). This has brought a renewed focus on other knowledge engineering tasks that are part of the knowledge graph construction process - such as modelling and consensus formation. In this talk, I argue that given the increasing performance of LLMs our attention should turn towards the choices made in the construction process and that value accrues to making good choices - or having good taste. I give concrete examples how to realise this in different stages of KG construction both by humans and with LLMs.

[1] https://sites.google.com/view/mehwish-alam/lecture-series-goblin , https://goblin-cost.eu/announcements/lunch-lecture-series/ 
[2] https://forms.gle/tLYJ2GeF9L6nvt7b7

— 
Best Regards
Mehwish Alam

Associate Professor
Telecom Paris, Institut Polytechnique de Paris
Department of Informatics and Networks (INFRES)
19 Pl. Marguerite Perey, 91120 Palaiseau, France

Web: https://sites.google.com/view/mehwish-alam/home