Research at the Department of Computer Science

 

The research at the Department of Computer Science is comprised of five intersecting areas: Bio- and Medical-Informatics, Vision and Cognition, Software and Systems, as well as Machine Learning and Theory. A commonality of all of these fields is the search for intelligent solutions that are for example based on machine learning techniques.
The Department of Computer Science as the sole technical department at the University of Tübingen is involved in the development of many fundamental and future oriented key technologies in close interdisciplinary cooperations. The applications range from advanced genome analyses to neural processing.

 

Computer Science in Tübingen is uniquely situated to take full advantage of research cooperation projects with renowned institutions in and around the area. The Max Planck Institutes for Intelligent Systems, Biological Cybernetics, and Developmental Biology, are at the core of these cooperations, adding to the department’s international visibility in the respective research fields. Moreover, close research collaborations take place with the Werner-Reichardt-Zentrum for Integrative Neurosciences (CIN), the Bernstein Center for Computational Neuroscience (BCCN), the Hertie Institute for Clinical Brain Research, and the Leibniz Institute for Wissensmedien.

 

The Department of Computer Science in cooperation with the Max Planck Institute for Intelligent Systems is the driving force behind the Tübingen part of the Cyber Valley initiative of Baden-Württemberg, which brings together key players from science and industry in the region to concentrate their research activities in the field of artificial intelligence and self-learning systems.


 

Research in Bioinformatics and Medical Informatics is rooted in the development of novel models, algorithms, and software tools dedicated to answering pertinent questions in the life sciences. Current and past research projects in Tübingen have dealt with developing solutions to challenging questions arising out of the diverse fields of phylogenetics, evolutionary of protein structures, structural bioinformatics, computational drug discovery, immuno-informatics, genomics, microbiome analysis, and gene expression analysis. Diverse and complex, these fields contain an immense opportunity for pushing the realm of knowledge and thereby solving contemporary problems in medicine, which require solutions to streamline the development of personalised cancer immune-therapies or close the gap in our understanding of microbiome-host interactions during an infection.

 

The Tübingen researchers in Bioinformatics and Medical Informatics are also members in the following interdisciplinary centers:

 

Currently funded research projects:

 

Working Groups:

 

Cooperation Partners:

  • Universitätsklinikum Tübingen
  • Max-Planck-Institut für Entwicklungsbiologie, Tübingen
  • Max-Planck-Institut für Menschheitsgeschichte, Jena
  • European Bioinformatics Institute, Cambridge, UK
  • Center for Bioinformatics and Computational Biology, University of Maryland at College Park, USA
  • Harvard Medical School, Boston, MA, USA
  • Dana Farber Cancer Institute, Boston, MA, USA
  • Life Sciences Institute, National University of Singapore,  Singapur
  • Biomathematics Research Centre, University of Canterbury,  Neuseeland
  • Australian Centre for Ancient DNA, Adelaide,  Australien


 

Intelligent systems are tranforming the world, be it in  industry, society or science. Advanced algorithms allow computers to autonomously drive cars,  they are used to design personalized medical treatments adapted to individual patients, or they can learn to play challenging strategy games such as Go. At the core of this revolution are methods and recent breakthroughs in the field of machine learning.
Machine learning is a subfield of computer science studying algorithms that extract structural information from data. Training directly on data, rather than being explicitly coded by human programmers, computers can learn to to recognize objects on images, to translate spoken langage into another language, or to classify different subtypes of cancer.
In our department we focus on understanding the basic princples behind machine learning:
Which conditions are necessary such that learning can take place,  what are its guiding principles? What kind of guarantees can we give on the results of machine learning algorithms? How can we quantify the uncertainty in our predictions? We also work on concrete applications of machine learning, for example in computer vision and robotics or  in medicine and biology. On a higher level, we work on enabling machine learning to take a more central role in the process of scientific discovery accross a wide range of scientific disciplines.

 

Research groups in our department:

  • Prof. Dr. Bjoern Andres
  • Prof. Dr. Andreas Geiger (Machine learning for computer vision)
  • Prof. Dr. Matthias Hein (Machine learning, from 1.5.2018)
  • Prof. Dr. Philipp Hennig (Machine learning methods, from 1.5.2018)
  • Prof. Dr. Ulrike Luxburg (Theory of Machine learning)
  • Dr. Mijung Park (Privacy Preserving Machine learning)
  • Prof. Dr. Nico Pfeifer (Machine learning in medicine)
  • Prof. Dr. Andreas Zell (Machine learning for robotics)

 

Associated professors: 

 

We closely collaborate with other machine learning researchers and institutes in Tuebingen, for example:

 

Jointly with our partners we run a number of activities:

 

 

The Software and Systems Engineering group is actively engaged in research that deals with the core practical and technical aspects of Computer Science from first principles to practical approaches. Applied projects are focused mainly on the development of large-scale software architectures, the analyses and transformation of structured data, algorithms for automatic evidence generation and optimization as well as the development of solutions for complex web-based distributed systems. Technical approaches are more focused on the analyses and optimization of complex embedded systems, and communication networks and computational architectures, all of which rely heavily on the application of machine learning algorithms.

 

Research topics:

Applied Computer Sciences

  • Declarative Computer Languages for the Analyses and Transformation of Structured Data
  • New Paradigms for Data-Intensive Programming
  • Efficient Construction of Large Software Systems
  • Definition, Analyses, and Verification of Software Systems
  • Algorithms for Calculation of Mathematically Conclusive logic

 

Technical Computer Sciences

  • Design and Verification of Secure Embedded Systems
  • Timing - and Power-Analyses of Embedded Systems
  • Architecture Design: From the Systems Level to Tapeout
  • Neural Interfaces and Brain Signal Decoding
  • Design, Optimization, and Application of Communication Networks
  • Software-Defined Networking, 5G and the Internet of Things

 

Research Projects:

  • BMBF-Projekte: autoSWIFT, EffektiV
  • DFG SPP 1500 Dependable Embedded Systems
  • DFG SPP 1835 Interaktiv Kooperierende Automobile
  • DFG ALIEN, CEPBCI
  • ECSEL THINGS2DO
  • Promotionskolleg Entwurf und Analyse Eingebetteter Systeme (EAES)
  • Zeiss Industry-on-Campus
  • Leibniz Gemeinschaft Wissenschaftscampus

 

Working Groups:

 

Cooperation Partners:

  • Bosch
  • Carl Zeiss
  • Daimler
  • IBM Deutschland Forschung und Entwicklung
  • Infineon Technologies
  • Siemens
  • Vector Informatik
  • Volkswagen

 

 

Theoretical computer sciences performs research on the very foundations of the field of computer science. It tries to answer fundamental questions such as: Which functions can be calculated by a computer in principle, and which ones cannot be computed? Which problems can be solved by efficient algorithms? How can we compare different algorithms, and in which sense are some algorithms "better" than others? Can computers "learn" to perform certain tasks, and if yes, which ones and how? Moreover, the field of theoretical computer science develops formal frameworks that can be used by other branches of computer science to describe and analyze complex systems. 
 

Research Topics in Tübingen:

  • Algorithmics: Our research topics cover a wide range of fields: combinatorics, discrete mathematics, graph theory, network analysis and visualisation, algorithmic geometry, computational social choice, parametric algorithms and algorithmic engineering. The goal of the latter is the practical implementation of theoretical methods.
  • Machine Learning: Our general research focus is the statistical analysis of machine learning algorithms, for example the statistical consistency of algorithms. Our particular focus is on unsupervised learning, in particular for data with "little mathematical structure" such as random networks or comparison-based data.
  • Verification: We work on model checking, temporal logic, SAT solving in theory and practice.
  • Formal languages and complexity theory: we work on the intersection between these two fields, one the one hand exploring the boundaries of computation, on the other hand developing concrete algorithms.

 

Working Groups:

 

 

The Department of Computer Science in Tuebingen is home to a unique interdisciplinary research group which has a strong focus on visual cognition, multi-sensor and sensor-motors processing and their interactions with the goal of making connections to the abstract underpinnings of human cognitive mechanisms and neural encoding processes. In order to achieve these research goals, which drive the understanding of psychophysics and cognitive neurosciences, sophisticated techniques in advanced image analysis, robotics, intelligent software systems, and computer graphics are employed.

 

Research Topics:      

  • Photo realistic 3D-Acquisition
  • Mobile Robots
  • Self-Learning Avatars in Virtual Reality Environments
  • Hand Eye Tracking
  • Computer Models of Human Vision
  • Generative, Recurrent, and Self-organized Artificial Neural Networks

 

Research Projects:

 

Working Groups:

 

Cooperation Partners: