mastodon 4.7.3Media Monitoring of the Past. Mining and connecting historical newspapers and radio. Led by EPFL DHLAB, ICL Uni Zurich, C2DH & UNIL. Funded by SNSF & FNR
Lausanne / Zurich / Luxembourg · impresso-project.ch
Posts currently synced from Twitter
📢 Registration is open for “The Avatars of Colonial Discourse: Computational Approaches to Colonial Media History.”
On 16 October 2026, the symposium will examine how computational methods can support research on colonial media archives, digitisation, infrastructures, silences, and the politics of knowledge.
📍 Belval, Luxembourg and online
🔗Register here: https://forms.gle/3q8XWzYtkoW5ng6h7
📻📰 The second Impresso II conference is a wrap! Last week in Lausanne, historians & digital humanists explored links between press and radio, from hybrid formats and careers to computational approaches.
📚 The book of abstracts is available online on Zenodo: https://doi.org/10.5281/zenodo.20813962
📢 Our third and last conference takes place in January 2027. CfP open: https://impresso-project.ch/news/2026/06/22/cfp-beyond-borders.html
📢 Join #Impresso as a Postdoctoral Researcher in Digital/Computational History at @C2DH, University of Luxembourg.
Work on data-driven historical research around “media influences” using multilingual newspaper and radio archives.
🗓️ Apply by 1 July 2026
📍 Belval Campus
💼 6 months | Full time
🔗 https://www.uni.lu/en/jobs/postdoctoral-researcher-in-digital-computational-history/
📢 Impresso at “AI through History, History through AI”
Tomorrow, 16 June, the Impresso team will lead a workshop at the University of Luxembourg on how AI-assisted retrieval shapes historical corpora. Using the Impresso Datalab, participants will explore multilingual and multimodal search, embeddings and entity-based methods, while assessing biases, omissions and interpretation.
Check the conference programme: https://www.uni.lu/c2dh-en/events/ai-through-history-history-through-ai/
How can historical newspapers and radio broadcasts 📰📻 from 500+ sources, 20+ institutions, multiple languages, and decades of digitization be represented for ML, information retrieval, and archival fidelity?
At PressMint / LREC2026, Pauline Conti, Simon Clematide, and Maud Ehrmann presented a framework now covering 567 titles, 73M+ content items, and 24B+ tokens 📊
🔗 Read the full paper: https://infoscience.epfl.ch/handle/20.500.14299/263437
📣 Call for participation: Impresso@BnF Datathon
📰 Historical Media Across Languages and Modalities
📍 Bibliothèque nationale de France, Paris
📅 28–30 Oct. 2026
Join us for a 3-day hands-on research event with the Impresso project and the Bibliothèque nationale de France. Work with historical media collections, the Impresso Web App, Datalab, NLP models, embeddings and case-study notebooks.
👉 Read more and register here: https://impresso-project.ch/news/2026/06/19/impresso-bnf-event-cfp.html
HIPE-2026 CLEF results are out! 🗞️📍
This third edition focused on person–place relation extraction in historical newspapers, with 17 teams and 45 runs.
🥇 Congratulations to the top teams across Accuracy, Generalization, and Accuracy-Efficiency profiles! 🎉
All competition data, submissions, and evaluation scripts are openly available:
📦 Evaluation repository: https://github.com/hipe-eval/hipe-2026-eval
📊 Results page: https://hipe-eval.github.io/HIPE-2026/results
🔎 Task details: https://hipe-eval.github.io/HIPE-2026/tasks
🔍 Impresso Seminar 2025 is coming up: a series of short online lunchtime sessions on current research in computational media history. Join us to explore new methods and findings in press history.
📅 Check the full program and register via the form on our website: https://impresso-project.ch/events/history/2025/10/28/Impresso-seminar-2025.html
(If you subscribe, please also check your spam folder for the confirmation email.)
📣 CALL FOR PAPERS: Radio and Newspapers: What Intersections for Media History?
The Impresso project is pleased to invite you to participate in its next conference, to be held in Lausanne in June 2026!
📆 The call is open until November 10, 2025
👉 Find out more here: https://impresso.github.io/radio-and-newspapers-conference/
🇵🇹 Impresso Datalab Hackathon | #DH2025 Lisbon
We’re excited to join the #DH2025 conference in Lisbon with a half-day Impresso Datalab Hackathon: Programmatic Access and Annotation Services for Multilingual and Multimodal Historical Media Collections.
📆 14 July 2025, 09:00–12:30
📍 NOVA-FCSH in Lisbon, Room B203 (TB)
We'll explore programmatic access, semantic enrichment, entity mapping, and notebook-based workflows for historical media 📰📻
📓 Full #DH2025 program: https://dh2025.adho.org/browse-the-program-agenda/
You've come a long way Baby 📄✨: The Impresso interface paper with @HistMethod is out! With focus on guiding theoretical principles, the integration of user requirements, semantic annotation layers and design + detailed usage examples 🔍 https://www.tandfonline.com/doi/full/10.1080/01615440.2024.2344004
🔮 Impresso at HistoFest 2025
This afternoon at HistoFest – History on Full Blast (C²DH), the Impresso team invites visitors to explore how witchcraft, superstitions and horoscopes circulated in the press over two centuries.
✨ Visit our stand to browse curious articles, try a fortune-telling mini app, and print your own historical horoscope!
📅 15 Oct 2025 – 16:30
📍 Belval Plaza, 16:30
🔗 Event: https://www.uni.lu/c2dh-en/events/histofest-history-on-full-blast/
📢 We’re hiring a Postdoc in Digital History! Join our Luxembourg-based team for a 21-month Postdoctoral Researcher position. If interested, or know someone who might be, you can learn more and find the link to apply here 👉 https://impresso-project.ch/news/2025/03/28/job-ad2.html
🚀 After two years in the making, we’re very excited to introduce a major step forward for the Impresso2 project !
🔹 Launch of the Impresso Datalab – programmatic access (API + Python) with notebooks for mapping, networks & enrichment.
🔹 Corpus expansion – 53 new titles, 180k+ issues, nearly 11M items (BnF, Swiss Economic Archives, BCUL, Swiss Federal Archives).
🔹 New data access management – fine-grained, partner-compliant user plans.
👉 Read the full release here: https://impresso-project.ch/app/search
Save the date! 📅 Nov 25-26, 2024: ONB Labs Symposium on "Newspapers as Datasets". The event will explore the potential of historical newspapers in Digital Humanities. Topics include AI, deep learning, and challenges in GLAM institutions. More info: https://labs.onb.ac.at/en/topic/symposium-2024
📢 We’re hiring a Postdoc in Digital/Computational History ! Join our Luxembourg-based team for a 28-month Postdoctoral Researcher position. If interested, or know someone who might be, you can learn more and find the link to apply here👉https://impresso-project.ch/news/2024/06/06/job-ad.html
The program for "Transmedia History: Circulations, Reconfigurations, and New Methodologies," co-organized by @ImpressoProject & the History Dept. of @unil, is out 👉 https://impresso.github.io/transmedia/
📅 Jan 27–28, 2025
📍 Lausanne, Switzerland
🔗Register here https://impresso.github.io/transmedia/registration
💡 Don't miss the "Digitale Methoden der Zeitungsanalyse" Summer School, Sept 11-12, 2024, at Zentralbibliothek Zürich! Register by Aug 11 👉 t.zbzuerich.ch/summerschool24. Highlights: Impresso project, digital image analysis, ANNO data research, German newspaper portal and more !
For more information, visit www.zb.uzh.ch or contact summerschool@zb.uzh.ch !
🔍Discover Impresso Project's use of ML in historical media! Check out @martenduering@fedihum.org’s lecture “Historische Medien und Maschinelles Lernen”, presented at the Digital History open research colloquium held at Humboldt University of Berlin .
Watch here: https://www.youtube.com/watch?v=7NL3yFHaQaQ
We were happy to participate in the 11th DH Benelux Conference in Leuven to discuss innovative methodologies for audio data 📻. Many thanks to Carolyn Birdsall for organizing the panel 'DH Research with Audio Data' and to participants Emily Hansell Clark & Loren Verreyen. #dhbenelux
🔥 Announcing HIPE-2026 at CLEF!
A shared task on extracting person–place relations and their temporal scope from multilingual historical texts.
Open to both LLM-based and traditional IE systems.
Two evaluation profiles:
🎯 Accuracy (balanced accuracy ranking)
🚀 Efficiency (accuracy + resource use)
Ideal for NLP, IE, and digital humanities researchers.
More info 👉 https://hipe-eval.github.io/HIPE-2026/
📢The Impresso team is hiring !
We’re looking for a Machine Learning / NLP Research Engineer to join us at the EPFL Digital Humanities Lab. You’ll help advance and consolidate NLP pipelines for large-scale, multilingual historical newspaper and radio archives—at the crossroads of research, engineering, and digital humanities.
🧩 Profile: 1–3 years ML/NLP experience (MSc/PhD)
📍 EPFL Lausanne
🗓️ Apply by 8 Dec 2025
💼 13-month full-time position
🔗 Details & application:
https://careers.epfl.ch/job/Lausanne-Machine-Learning-Research-Engineer-in-Natural-Language-Processing-and-Media-Mining/1163726355/
📚 Impresso took part in the Assises thématiques sur l’IA by the #MinistèreCultureLU at 1535° in Differdange. We ran a workshop on using the Impresso Web App and Datalab to explore digitised historical media. Participants tested semantic enrichments, multilingual search and notebooks. We also previewed Impresso Barista 🤖☕ and our multilingual embeddings for cross-lingual analysis.
🔗Impresso Web App: https://impresso-project.ch/app/
🔗Impresso Datalab: https://impresso-project.ch/datalab/
Impresso 2 at #DigiHistCH2024: We presented our work connecting digitised newspapers & radio across Western Europe using NLP & advanced semantic layers. Exciting conversations on the future of digital tools in history! Read more: https://digihistch24.github.io/book-of-abstracts/submissions/443/
📢Do you work with sources from different types of media? We'd like to remind you that you only have a few days left to respond to the call for papers for our “Transmedia History" conference: deadline is July 15! All the info here 👉 https://impresso.github.io/transmedia/
📰 Can LLMs help correct historical OCR?
Tomorrow, Impresso presents its work on LLM-based OCR post-correction at the CENL “AI in Libraries” webinar, alongside Mistral AI, the Bibliothèque nationale de France, and the British Library.
📅 Friday 10 July 2026, 10:00 Paris time
🔗 Registration: https://www.cenl.org/network-group-ai-in-libraries-webinars-2026/
#OCRpostcorrection #LLM #NLP #HistoricalDocuments #ImpressoProject
📰 Recap: yesterday, the Impresso team ran a workshop at “AI through History, History through AI” in Belval with Vy Cao, Caio Mello, Marten Düring, Daniele Guido & Ferdaous Affan.
With the Impresso Datalab, participants explored AI-assisted retrieval through two cases: Mega Events, presented by Caio Mello, and Medicinal Catalogues, presented by Vy Cao, followed by a brainstorming session on embeddings and corpus building.
https://impresso-project.ch/app/
#Impresso #DigitalHistory #AI
🏆 Our paper “Data Augmentation for Historical NER” received the Best Paper Award at #SwissText2026, organised by the Department of Computational Linguistics at UZH and held at the University of Zurich.
We compare lexical mention replacement with LLM-based corpus annotation for historical newspapers. Simple mention replacement delivers consistent gains across French and German.
📄 https://aclanthology.org/2026.swisstext-1.14/
💻 https://github.com/impresso/impresso-named-entity-data-augmentation
📣 We are happy to announce another opportunity to join the Luxembourg-based Impresso team.
Applications are open for 3 fully funded Visiting Fellowships at C2DH, University of Luxembourg, for scholars at postdoc or professor level working with Impresso data and the Impresso Datalab.
📍 Luxembourg
⏳ 3 months
💶 Travel, housing & daily allowance covered
📅 Deadline: 1 July 2026
Read more and apply here: https://impresso-project.ch/news/2026/06/09/fellowships-ad.html
📢 Impresso at #DHBenelux2026
Tomorrow, 2 June, the Impresso team will run a hands-on workshop in Maastricht on “Embedding Newspapers for Multimodal and Multilingual Data Analysis with the Impresso Datalab.”
📍 Maastricht University
🕑 14:00–17:00
We’ll explore AI-assisted methods, embeddings, metadata, image captions and semantic enrichments for historical media analysis.
