Categories
Presentations Updates

Jiaqi Presents Semantic Change Research at CHR 2025 in Luxembourg

In December, TRIFECTA team member Jiaqi travelled to Luxembourg to attend and present at CHR 2025, the 6th Conference on Computational Humanities Research, held from December 9-12 at the Luxembourg Centre for Contemporary and Digital History (C²DH) at the University of Luxembourg.

A Venue Where Past Meets Future

The conference took place on the University’s Belval campus. There is something poetic about discussing historical language change here, given the campus’s own remarkable transformation. Where steelworkers once tended roaring blast furnaces that helped build Europe’s railways, researchers now gather in sleek modern buildings to study the past with computational tools.

The campus sits in the heart of what was once Luxembourg’s industrial heartland. Two massive blast furnaces still tower over the site, preserved as monuments to the region’s steel-producing heritage. The university buildings have been designed to echo this history: the 85-metre Maison du Savoir deliberately mirrors the proportions of the old furnaces, while the Maison du Livre wraps around a former ore silo. Walking between sessions, you pass reflective water basins that occupy spaces where molten steel once flowed. It’s a striking reminder that transformation—whether of industrial sites or of word meanings—is a constant in human history.

Tracing How Words Change

Jiaqi presented this paper on semantic change detection in historical Dutch newspapers, exploring how computational methods can help us understand how the meanings of words evolve over time. This work is in line with TRIFECTA’s mission of developing better knowledge graphs for humanities research, as understanding how concepts shift and transform across historical periods is crucial for accurately representing historical knowledge.

Connections and Conversations

CHR has always been a warm and welcoming community for researchers working at the intersection of computation and the humanities, and this year was no exception. The conference brought together scholars working on everything from historical NLP to cultural analytics, and the conversations over coffee were just as valuable as the formal sessions. It was particularly inspiring to connect with others grappling with similar challenges around temporal analysis and historical languages.

We look forward to building on these connections and continuing to advance our research for the TRIFECTA project. We look forward to CHR2027 (the next edition) in Manchester! 

Categories
Achievements Presentations Publication

TRIFECTA Team Takes Home Two Awards @LDK

Gauri accepting the Best Student Paper Award from conference chair Andon Tchechmedjiev

In September, team members Gauri, Jiaqi, Marieke, Rik, and Teresa travelled to Naples, Italy to attend and present at LDK 2025, the 5th Conference on Language, Data and Knowledge. This conference is very central to the different research strands in the project, so it was great to (re)connect with colleagues and see what they are working on (and eat great food).

We presented the following papers:

  • Veruska Zamborlini, Jiaqi Zhu, Marieke van Erp, and Arianna Betti. Philosophising Lexical Meaning as an OntoLex-Lemon Extension. (presented at the satellite OntoLex workshop). This research is part of our knowledge modelling strand and in this paper we investigated how we can represent different aspects and meanings of a concept through time or in different contexts;
  • Gauri Bhagwat, Marieke van Erp, Teresa Paccosi, Rik Hoekstra. Detecting Changing Culinary Trends Through Historical Recipes. This research is part of our food history use case, and presents an analysis of different editions of a cookbook as well as newspaper recipes to see how ingredient use changes over time;
  • Marieke van Erp, Jiaqi Zhu, Vera Provatorova. Tracing Organisation Evolution in Wikidata. This paper is an investigation of how change is represented in one of the largest and most commonly used knowledge graphs. As we are considering feeding any data generated within the project back into this, it is necessary to know if existing data models are a suitable fit;
  • Andrea Schimmenti, Stefano De Giorgis, Fabio Vitali, Marieke van Erp. Old Reviews, New Aspects: Aspect Based Sentiment Analysis and Entity Typing for Book Reviews with LLMs. In this collaboration with the University of Bologna and the Italian National Research Council, we investigated the use of large language models to analyse opinions in a data-scarce domain. Whilst we used a different use case domain than TRIFECTA’s main maritime and food history use cases, we think it is important to see what connections to other domains we have and how tools work there to see if we can translate them to our use cases.

While it’s already great to get papers accepted to a conference and present and discuss them with colleagues, it was even cooler to see our efforts recognised by the fact that Detecting Changing Culinary Trends Through Historical Recipes coordinated by Gauri won the Best Student Paper Award and our Tracing Organisation Evaluation in Wikidata paper won the Best Poster Award!

The winning poster