Workshop 7

Workshop 7: Machine Learning to Read Yesterday’s News. How semantic enrichments enhance the study of digitised historical newspapers

Verantwortliche: Marten Düring, Estelle Bunout
Organisation:Luxembourg Centre for Contemporary and Digital History, Luxemburg
Datum: Mittwoch, 24.05.2023, 9-13 Uhr
Ort:Raum 2070 A, Humboldt-Universität zu Berlin, Hauptgebäude, 1. Obergeschoss, Unter den Linden 6, 10117 Berlin


Our goal for this workshop is first for historians to familiarise themselves with the opportunities and pitfalls of semantically enriched sources. Second, we will introduce participants to the complexity of digitised newspaper collections and focus on key operations for their exploration and analysis. Third, we will encourage participants to discuss the capacities of the impresso app to support historical research and to offer transparency regarding data processing and interface functions. 

Technische Voraussetzungen

Die Teilnehmer:innen werden gebeten, nach Möglichkeit einen Laptop mitzubringen.


HistorikerInnen / GeisteswissenschafterInnen, keine weiteren Voraussetzungen nötig

Info-/ Ankündigungstext

In this workshop we will use the impresso app to explore opportunities and challenges which accompany the semantic enrichment of historical newspapers. We will reflect on the added value of Natural Language Processing techniques such as topic modelling, text reuse detection and word embeddings for historians in conjunction with an introduction and critical assessment of design solutions for the scalable reading of such enriched sources. We target researchers at all (digital) skill levels.

During the workshop participants will address generic, yet complex historical questions such as: “How did the news about the Titanic catastrophe travel through the media sphere?” or “What constitutes a “crisis”? And why did it peak in 1932 in the press?”, and thereby explore the potential of the impresso app to assist historical research workflows: collection creation, query comparison, content filtering based on topic models, named entities (persons and places), article recommendations, image similarity, ngrams and text-reuse clusters. 

The final part of the workshop will be dedicated to discussing the capacities of the impresso app to offer transparency regarding data processing and interface functions.

Participants will be able to take home knowledge of advanced usage of the impresso interface as well as an awareness of the peculiarities of semantically enriched historical sources. 

Die Pre-Conference-Workshops werden im Rahmen der diesjährigen Digital History-Tagung angeboten. Die Anmeldung zu den Workshops erfolgt zusammen mit der Tagungsanmeldung über ConfTool.

Zurück zur Programmübersicht

Suche in OpenEdition Search

Sie werden weitergeleitet zur OpenEdition Search