Max Planck Institute for Software Systems, Saarbrücken site

With twelve Synergy Grants, the Max Planck Society claims top spot in the ERC ranking

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

A new technology allows software components to be isolated from each other with little computational work

ERC awards Advanced Grants worth up to 2.5 million euros each.

The Max Planck Society's 67th Annual Meeting wrapped up with a panel discussion the "Internet of Things"

When he was still in primary school, Krishna Gummadi learned to play musical instruments and studied programming. He soon gave up on music, but programming turned out to be his calling. These days, as director at the Max Planck Institute for Software Systems in Saarbrücken, he is researching, among other things, why artificial intelligence often makes decisions that are just as discriminatory as the ones humans make, and how this can be prevented.

Artificial intelligence has the power to support people in an ever-increasing number of areas – including education. Researchers at the Max Planck Institute for Software Systems are working under Adish Singla’s leadership to find methods to help children learn how to program. These algorithms can, however, also be used in other areas.

He describes himself as a bit unconventional. Derek Dreyer does, indeed, work with languages. Not, however, with everyday languages, but with programming languages, since they’re more logical. The U.S.-born researcher works at the Max Planck Institute for Software Systems in Saarbruecken. Even though science is an important part of his life, he still makes room for passions like music. And then there’s the thing about whisky – unconventional indeed.

We have barely any control over where information about us and even photos bearing our likeness are displayed. In the future, however, it may at least be possible to prevent ourselves from appearing as bystanders in photos on other people’s Facebook pages. This is thanks to technology developed by a team working under Paarijaat Aditya, Rijurekha Sen and Peter Druschel from the Max Planck Institute for Software Systems.

When a computer takes forever to load a website, it may be annoying, but it is nothing more serious than that. If, however, the electronics in a car or a plane don’t process commands exactly when they are supposed to, the consequences can be fatal. Björn Brandenburg and his team at the Max Planck Institute for Software Systems in Kaiserslautern and Saarbrücken study how to construct real-time systems in such a way that it can be proven that they always react on time.

We introduce the algorithmic problem of determining whether the solution to a given linear differential equation has a zero or not (the Zero Problem). Somewhat surprisingly, this is a longstanding open problem, with deep connections to several other areas of mathematics. We report on some of our recent progress in this area.

A fundamental societal challenge is to provide cost-effective and inclusive education that fosters general problem solving skills. This calls for the development of next-generation AI-driven educational technology for open-ended conceptual domains.

Many of the computer programs we use today are distributed systems running in datacenters. For people who run and maintain these programs, diagnosing problems can be very challenging, because symptoms and root causes can be spread across many different machines and system components. In order to make it easier to operate complex systems, we need tools that record and analyze their runtime behavior, and to localize and diagnose problems when they Distributed tracing tools piece together what happened at runtime, and can help localize and diagnose problems when they occur.

Enabling engineers and programmers to automatically verify the correctness of the computer systems that they design is one of the Grand Challenges of computer science. The Foundations of Algorithmic Verification group focuses on a series of fundamental algorithmic problems for dynamical systems, with the overarching goal of offering a systematic exact computational treatment of various important classes of such systems and other fundamental models used in mathematics, computer science, and the quantitative sciences.

Residential broadband ISPs like DSL & cable are being used by hundreds of millions of people to access the Internet. Today, ISPs are deploying firewalls and traffic shapers to manipulate the performance of user traffic. Most ISPs do not reveal the details of their networks to their customers, preventing them from making an informed choice of their ISP. The goal of our Glasnost project is to make access networks more transparent to their customers. To date, more than a hundred thousand users world-wide used Glasnost to test if their ISPs are interfering with their peer-to-peer traffic.