Multilinguality and Language Technology

The research department Multilinguality and Language Technology (MLT) in Saarbrücken is headed by Prof. Dr. Josef van Genabith.

MLT covers the following main topics: Machine Translation (MT), Question Answering and Information Extraction Systems (QAIE), Talking Robots (TR), and Data and Resources (D&R).

  • MT group: Machine translation between texts and from text to sign language;
  • QAIE group: Highly scalable text analysis and, based on this, question-answering and information extraction systems;
  • TR group: Dialog processing for human-robot collaboration and for teamwork support. Basic dialog technologies as well as applications, e.g., for disaster response management and health care;
  • D&R group: Collection of language data and processing into resources for the development of intelligent language technologies.

We work with machine learning and knowledge-based approaches. Besides basic research, MLT is strongly involved in contract research and industrial projects. 

At Saarland University, members of the MLT group are involved in leading positions in SFB 1102 (Information Density and Linguistic Encoding, subproject B6), a DFG project on multi-modal interfaces for post-editing of MT results (MMPE), and the “Erasmus Mundus European Masters Program in Language and Communication Technologies” (LCT).

TRADR
ELRC
QT21

Talks at MLT-Meetings

2022

Presenter Affiliation Title Date
Daria Fedorova DFKI Saarbrücken Cross-Lingual Knowledge Transfer for Sequence Labeling Tasks Dec 20
Tanja Bäumel DFKI Saarbrücken Feature Textualization: Using Activation Maximization to reveal linguistic knowledge in Pretrained Language Models Dec 6
Cristina España i Bonet DFKI Saarbrücken The (Undesired) Attenuation of Human Biases by Multilinguality. Nov 29
Cennet Oguz DFKI Saarbrücken Chop and Change: Anaphora Resolution in Instructional Cooking Videos Nov 22
Natalia Skachkova, Qiankun Zheng DFKI Saarbrücken Title 1: Anaphora Resolution in Dialogue: 3 systems submitted by DFKI/Inria team for CODI-CRAC 2022 Title 2: A transformer-based span classification model. Shared Task. Nov 15
Natalia Skachkova DFKI Saarbrücken Enriching contextual embeddings with semantic knowledge: Intermediate project results and current challenges Nov 8
Lukas Lange Bosch Center for Artificial Intelligence Robust Input Representations for Low-Resource Information Extraction Oct 25
Thierry Declerck DFKI Saarbrücken Towards a RDF Representation of the Infrastructure consisting in using WordNet(s) as a conceptual Interlingua between multilingual Sign Language Datasets Oct 18
David Lindemann University of the Basque Country Wikidata: From Concepts to Lexemes Oct 10
Natalia Skachkova DFKI Saarbrücken Chinese Event Discourse Deixis Resolution: Design of the Dataset and Model Oct 4
Eric Rietzke DFKI Trier Vorstellung von SPELL – KI in der Leitstelle der Zukunft Sep 29

Contact

Office:
Phone: +49 681 85775 5292

Deutsches Forschungszentrum für Künstliche Intelligenz GmbH (DFKI)
Campus D3 2
Stuhlsatzenhausweg 3
66123 Saarbrücken
Germany

News from the Research Department

[Translate to English:]

Through an urgent plea for language data, the European Commission (EC) is supporting the development of technologies to enable cross-border…

Go to Article

Together with the German Research Center for Artificial Intelligence (DFKI), the Informationsdienst Wissenschaft (idw) has developed, evaluated, and…

Go to Article

On 28 July 2021, Andrea Lösch and Thierry Declerck from the research group "Multilinguality and Speech Technology" were invited to give a talk about…

Go to Article

German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz