KEEPHA Partners

The German Research Center for Artificial Intelligence (DFKI) is Germany's leading business-oriented research institution in the field of innovative software technologies based on artificial intelligence methods. The Speech and Language Technology Lab (SLT-Lab) is a worldwide recognized research group for natural language processing with various focal points in all important areas of language technology, such as information extraction, big text data analysis and semantic search as well as biomedical AI research. Prof. Dr.-Ing. Sebastian Möller is the head of the Speech and Language Technology Lab and professor at the Technical University of Berlin. His research group at DFKI has a strong background in corpus generation (Cotik et al. 2017, Schiersch et al., 2018), general information extraction (Alt et al., 2019a, Alt et al. 2019b), as well as processing biomedical text (Kirschnick et al., 2018, Roller et al., 2018). It researches neural information- and relation-extraction with regard to explanatory methods and continuous learning in neural networks as well as scalable learning for medical prognosis models. The SLT-Lab is a member of the Berlin Big Data Center (BBDC) and is involved in the projects DeepLee, MACSS, and BigMedilytics.

Nara Institute of Science Technology (NAIST) is a Japanese government-funded University, specialized for priority scientific areas, Information Science, Biological Science and Materials Science, having only graduate schools. The Computational Linguistics Lab is a world-wide recognized research group that has implemented and publicized a number of natural language processing tools such as ChaSen, MeCab and CaboCha, the de-facto standard processing tools for Japanese. Yuji Matsumoto, the head of the Lab, is the principal investigator of JST CREST project of scientific document analysis. His group has produced a number of document and text analysis tools for extracting knowledge from scholarly documents. He is also leading the Knowledge Acquisition Team at RIKEN Center for Advanced Intelligence Project (AIP). Assistant Professor Hiroyuki Shindo at NAIST plays a major role in developing a number of document analysis and knowledge extraction systems. The major collaborator, Professor Akiko Aizawa at National Institute of Informatics (NII) is one of the most active researchers in text mining and information extraction in Japan. One of her major research topics is logical and semantic structure analysis of research papers for scientific knowledge acquisition.

The Computer Science Laboratory for Mechanics and Engineering Sciences of the French National Centre for Scientific Research (LIMSI/CNRS) accommodates 120 permanent employees. It is one of France's largest research laboratories working on language technologies; it covers the full spectrum from low level signal processing to natural language processing, text mining and machine translation. Faculty members in this project are P Zweigenbaum, PhD, Senior Researcher, former head of LIMSI's ILES NLP team, T Lavergne, PhD, Adjunct Professor, C Grouin, PhD, Research Engineer, A Névéol, PhD, Researcher, S Ghannay, PhD, Adjunct Professor, T Hamon, PhD, Adjunct Professor, and P Paroubek, PhD, Research Engineer. Particularly relevant to this project, LIMSI brings a strong expertise in information extraction (all cited members) in the biomedical domain (PZ, CG, AN, TL, TH, PP), scientific literature mining (PP, AN, CG, TH, PZ) and patient forum mining (CG, PZ), biomedical computational terminology (TH, PZ, AN, CG) and knowledge representation (PZ), word embeddings (SG, TL, PZ), multilingual language processing (TL, PZ, TH), corpus annotation (CG, AN, PP, TH, PZ, TL), and NLP shared tasks and evaluation (PP, CG, TH, AN, PZ). TL is specialized in machine learning for information extraction, and is the author of the Wapiti CRF toolbox. AN, PP and PZ participated in the MiRoR H2020 International Training Network in which they supervised students and organized training actions. LIMSI's most relevant recent projects include ANR ADDICTE on improved word embeddings for specialized domains, and Vigi4med and PHARES on information extraction from patient forums for pharmacovigilance, in which LIMSI delivered an information extraction pipeline to the French National Agency for the Safety of Medicines (ANSM).