ITI - Institute of Theoretical Informatics
The Institute of Theoretical Informatics deals with the theoretical fundamentals of informatics and their application to practical issues. Its wide range of subjects covers algorithmics, formal methods, cryptography, and security.
The institute, among other things, focuses on algorithmics and its applications, in particular on graph algorithms, algorithmic geometry, and parallel and distributed algorithms with special emphasis on algorithm engineering methodology. The research subjects comprise theoretical and practical issues from different areas e.g., algorithm libraries for basic algorithms and data structures, network analysis and visualization, and algorithmic methods for traffic and energy system optimization.
Besides, ITI focuses on the development of algorithms and models for solving problems in the field of evolutionary bioinformatics. The relevant research subjects comprise classical sequence analysis, algorithms and software for reconstructions of phylogenetic trees, population genetics methods, and analyses of very large biological data sets on clusters and supercomputers.
The algorithms groups participate in the DFG Research Training Group GRK-2153 »Energy Status Data - Informatics Methods for its Collection, Analysis and Exploitation«, the DFG Priority Program SPP-1736 »Algorithms for Big Data«, the DFG Research Group FOR-2083 »Integrated Planning in Public Transport« and the Helmholtz Programs »Storage and Cross-linked Infrastructures« and »Supercomputing & Big Data«.
The research group Artificial intelligence for materials science (AiMat) works on the development and application of AI and machine learning (ML) methods for materials science and chemistry. Focus topics are (graph) neural networks for the prediction of molecular and materials properties, the combination of interpretable AI methods, ML-enhanced simulations and automated experiments to autonomous materials acceleration platforms, and generative models for inverse materials design.
The group AI for Climate and Environmental Sciences (KI-Klima) combines machine learning/AI, numerical models and Earth observations to address key challenges in the climate and environmental sciences.
For example, using satellite data, the group develops novel machine learning approaches to reduce uncertainties in global climate change projections, to speed up and improve global Earth system models run on supercomputers, and uses machine learning to improve our understanding of the complex, highly-coupled climate system.
In particular, the group is working on data-driven causal discovery and explainable AI.
In addition, machine learning techniques are developed for better and more affordable measurements of the Earth system.
Besides theoretical fundamental research e.g., in the fields of semantics of programming languages or development of suitable calculi, the area of formal methods comprises the development of formal-logic methods for practice through integration with conventional software engineering methods and application of formal methods to improve the reliability of software in concrete areas of application.
The efficient and user-friendly specification, verification, and automatic analysis of software, both at the abstract design level and implementation level, is another important subject.
One focus is on the verification of object-oriented software. The respective research activities, among other things, are part of the KeY project (www.key-project.org) which is aimed at the deductive verification of Java programs.
Software safety, correctness of microkernels, network protocols, numerical software, and choice methods are among the present fields of application.
Another ITI focus is on cryptography and security research. In 1988, a correspondent study group was founded as European Institute for System Security (EISS). Today, EISS participates in interdisciplinary research at the Competence Center for Applied IT Security Technology (KASTEL), carries out research transfer via Forschungszentrum Informatik (FZI), and takes part in the Helmholtz Program “Supercomputing and Big Data”.
All research activities are spanned by the central topic of “provable security guarantees”. These are formal proofs that systems within mathematical models cannot be attacked. Proofs of that kind provide traceable and reliable security guarantees and allow mathematical access where security is proved mathematically based on falsifiable assumptions.