The GDR TAL 2024 Day will take place at L3i on November 7, 2024, in La Rochelle, on the theme "Natural Language Processing and Digital Humanities" with the support of the Région Nouvelle-Aquitaine ANNA project (2019-1R40226). The day will be organized around invited oral presentations, poster presentations, demos (see the call for submissions below), and a round table discussion.
Natural language processing (NLP) methods and tools have gradually taken on an increasingly significant role in digital humanities (DH). Numerous disciplines in the humanities and social sciences regularly rely on NLP to address various issues in heterogeneous chronological, documentary, linguistic, or stylistic contexts. Adapting NLP methods to the research objectives of DH raises many challenges, which we propose to discuss from an interdisciplinary perspective.
The objective is to bring together researchers specializing in NLP (both written and spoken) and disciplines from DH to discuss issues, recent advances, and research perspectives at the intersection of NLP and DH.
Topics (non-exhaustive list):
Corpus creation (texts, speech) for digital humanities
Automatic annotation of ancient documents (handwritten texts, archives, etc.)
Information extraction (texts, speech, historical documents)
Adapting NLP interfaces to the specific needs of digital humanities researchers
User modeling for resource recommendation systems (texts, archives, manuscripts)
Conversational agents applied to digital humanities
Development of Q&A systems
Fact-checking in documentary corpora
Explainability of classification or text categorization choices
Transparency in models used for corpus analysis in digital humanities
Interactive text summarization for digital humanities
Summary generation
Exploration of multimodal corpora (texts, speech, images, videos) and geolocated content for digital humanities
Multimodal interaction for information retrieval
LLM and digital humanities
Robust analysis methods for digital humanities
As part of this day, we invite researchers working on these topics, in both academic and industrial contexts, to present their work (demo or poster), even if already published, to exchange with colleagues in the field. To participate, simply submit a one-page abstract and/or the poster if it already exists and/or the article describing the work if already published in French or English.
Submissions should be made on the Sciencesconf website or sent directly to the organizers.