Roland Roller


I am a researcher and project manager in the Text Analytics (LT) group at the German Research Center for Artificial Intelligence (DFKI). I focus on topics related to natural language processing and machine learning with a high interest in the biomedical and clinical domain. Currently I am working within the research projects MACSS ( Medical Allround-Care Service Solutions) and BigMedilytics (Big Data for Medical Analytics).

The last years, me and my collegues were developing resources and tools to process German medical text. We created various datasets and trained models for syntactic analysis and to extract relevant medical information from clinical text. Within the MACSS project we developed methods and models to cover the complete NLP pipeline to be able to process text of the nephrology domain. If possible we try to make our work (data and models) available for the community.
Currently I am focussing particularly on methods to support clinicians in their daily routine. In BigMedilytics we develop methods to detect high risk patients in context of kidney transplants. We work with longitudinal data including information such as demographics, vital parameter, lab values, medications (adherence) and clinical notes.

Beside that I have got a high interest in the generation, exploration and usage of knowledge graphs and biomedical ontologies to support NLP tasks, as well as the analysis of social media in context of health related topics.


22.11.2018How beer-related tweeting climaxed during the World Cup in Technology Review
12.11.2018Talk together with Wiebke Düttmann-Rehnolt at European Big Data Value Forum (EBDVF) 2018 in Vienna: "Optimizing the care of kidney transplant patients using Big Data"
7.11.2018Talk at Future Medicine 2018 in Berlin: "Early detection of complications in patients after kidney transplant"
16.10.2018Talk at Smart Data Forum in Berlin: "Medical Data Analytics for Chronic Kidney Disease"
8.10.2018Talk at Berlin-Brandenburgische Akademie der Wissenschaften in context of Medizindiagnostik und KI: "Informationszugriff und Medizindiagnostik in der Nephrologie"
October 2018Accepted paper "mEx - An Information Extraction Platform for German Medical Text" for Semantic Web Applications and Tools for Healthcare and Life Sciences (SWAT4HCLS'2018)
August 2018Accepted paper "Football and Beer - a Social Media Analysis on Twitter in Context of the FIFA Football World Cup 2018" for SMM4H 2018
July 2018Accepted paper "A Domain-adapted Dependency Parser for German Clinical Text" for KONVENS 2018
June 2018Our book article "Integrierte Versorgung chronisch kranker Patienten am Beispiel von MACSS" is published as open access in IT für soziale Inklusion
11.06.2018Language Technology Seminar at TU Berlin: Knowledge Graphs
08.05.2018Talk at MultilingualBIO in Miyazaki: Cross-lingual Candidate Search for Biomedical Concept Normalization
30.04.2018Language Technology Seminar at TU Berlin: Data & Language Resources
April 2018Article Ärzteblatt
23.03.2018Panel Discussion XPOMET in Leibzig: "Healthcare for all: The transformative potential of artificial intelligence"
6.-7.03.2018BigMedilytics KickOff in Eindhoven
FebruaryAccepted paper at MultilingualBIO in Miyazaki
29.01.2018MACSS Milestone Meeting in Berlin