Max Planck Institute for Informatics

Many publications by Max Planck scientists in 2023 were of great social relevance or met with a great media response. We have selected 12 articles to present you with an overview of some noteworthy research of the year

Special shapes can release active substances in a controlled manner

Funny looking pills are not a design gimmick, they can release medication in a desired time regime!

Max Planck Institute for Informatics and Google engage in a strategic research partnership

The cooperation strengthens application-related research on artificial intelligence in Germany

The first self-driving cars are already on the road. Yet the technology is not yet fully developed, and certain ethical issues remain unsolved. It is also high time to think about how the new technology can promote sustainable transportation in the future.

Logical processes help computers crack complex mathematical problems, thus making them smarter and faster. Logic can even take human-machine communications to a whole new level. Christoph Weidenbach from the Max Planck Institute for Informatics has been developing such promising logical algorithms for thirty years, and is even testing them on his own race car.

Our eyes are our window on the world, but they also reveal a lot about us. Andreas Bulling and his team at the Max Planck Institute for Informatics in Saarbruecken and at the University of Stuttgart exploit this by teaching computers to interpret our gaze. Ultimately, their aim is to enable robots or avatars to communicate with us eye to eye.

Today, the Internet is just as much a part of our everyday life as our refrigerator. Yet researchers are constantly working to develop it further so that it can continue to function despite all the innovations. Anja Feldmann, Director at the Max Planck Institute for Informatics in Saarbruecken, is one of them. However, there are some problems that she can only solve when she focuses her entire attention on her hobby.

He was one of the first computer science students in Germany. Today, Kurt Mehlhorn, Director at the Max Planck Institute for Informatics in Saarbruecken, can look back on the many problems he has solved – with solutions that are also applicable to navigation systems and search engines. At least as important to him, however, are the numerous academic careers that began in his group. And he still has ideas for new research projects.

Virtual worlds are a central component for visual communication in the future. But how can these worlds be built so that they look photorealistic and can be explored smoothly? How do we create models that are flexible enough so that users can actively influence them? These questions targeting the fundamentals of computer graphics are being investigated at the MPI for Informatics. The application of artificial intelligence also plays a crucial role in this process.

Along with the inventions of new computing devices, a growing number of new computing paradigms emerges that lead to new challenges in designing computer algorithms. Researchers at MPI for Informatics see these challenges as a new way to think about algorithms and develop a research program that provides answers to questions that computer scientists could not figure out for decades and at the same time provides new generations of network algorithms for future computing devices and data.

The optical data center network is a cutting-edge technology to build network infrastructure for cloud computing. It has great potential as well as challenges to adoption. At the MPI for Informatics, we develop practical ways of implementing optical data center networks to serve ever-increasing user demands for cloud applications in the future.

Learning new insights is a key technology in solving hard problems with a computer. The systematic generation of new insights also implies the forgetting of insights in order to meet space and time restrictions.

Next-generation search engines should understand quantitative measures of entities. For example, it should be easily possible to retrieve all entrepreneurs worth between 1 and 10 billion Euros, and a query for electric cars with CO2 footprint below 5kg/100km should be answered precisely and concisely. We are developing algorithmic and machine-learning methods towards this goal. Our Qsearch prototype supports searching for financial, technical and physical quantities. Ultimately, we aim to support analysts in effortlessly tapping Internet contents for advanced studies.