People increasingly use artificial intelligence today in important decisions that have a major impact on people’s lives – from improving medical diagnosis and increasing efficiency in research to supporting data-based decision-making in business and administration. However, AI cannot ‘understand’ like a human being why something happens or how damage could be avoided by acting differently, for example when driving an autonomous vehicle or in robotics. ‘Many real-world systems today are based on AI, but their actions are difficult to interpret because we often don’t know exactly what the neural networks are calculating in the background. This means that IT systems are not particularly reliable and are difficult to manage when they have to comply with safety regulations, for example,” says Sara Magliacane.

The computer scientist therefore wants to research at Saarland University how ideas of causality can be used to make complex AI models safer and more reliable. Among other things, the focus is on the large language models on which ChatGPT and similar programs are based. It will also look at so-called vision language models, in which AI is trained to link text data and visual data such as images and videos. Another focus will be on ’embodied AI,’ which can be used to program care robots, for example, to ‘understand’ information from their environment and translate it into their own actions.

Sara Magliacane’s areas of focus enrich the research environment at the Saarland Informatics Campus in many ways. They are particularly well suited to the Research Training Group ‘Neuroexplicit Models for Language Processing, Image Recognition and Action Decisions’, in which several computer science research institutes in the region are involved alongside Saarland University and which will continue to receive funding from the German Research Foundation until 2028.

Short CV of Sara Magliacane 

Sara Magliacane has been an assistant professor at the Amsterdam Machine Learning Lab at the University of Amsterdam since 2020. From 2019 to 2025, she was also a researcher at the renowned MIT-IBM Watson AI Lab in Cambridge, Massachusetts (USA), where she was IBM’s principal investigator for several projects in collaboration with the MIT faculty. Born in Italy, she completed her master’s degree at the Polytechnic University of Milan (Italy) and received her Ph.D. from the Free University of Amsterdam (Netherlands) in 2017. She then spent two years conducting research at the IBM Thomas J. Watson Research Centre before moving to the MIT-IBM Watson AI Lab and the University of Amsterdam.

The Harlan D. Mills Award is presented by the international engineering association IEEE (Institute of Electrical and Electronics Engineers) and recognizes long-standing and impactful research contributions to software development. Andreas Zeller is being honored for his lasting contributions to software debugging, program analysis, mining software repositories, and automated test generation.

The Harlan D. Mills Award will be presented to Andreas Zeller in April at the International Conference on Software Engineering in Rio de Janeiro. IEEE has more than 500,000 members from over 190 countries. It publishes about one third of the world’s technical literature in electrical engineering, computer science, and electronics, and supports more than 2,000 conferences each year.

Background Saarland Informatics Campus:
900 scientists (including 400 PhD students) and about 2,500 students from more than 80 nations make the Saarland Informatics Campus (SIC) one of the leading locations for computer science in Europe. Four world-renowned research institutes cover the entire spectrum of computer science, namely the German Research Center for Artificial Intelligence (DFKI), the Max Planck Institute for Informatics, the Max Planck Institute for Software Systems, the Center for Bioinformatics as well as Saarland University with three departments and 24 degree programs.

This text was machine translated from the German with no human editing.

They published their findings in the renowned scientific journal Nature Communications.

“Flights to the International Space Station ISS place a strain on astronauts in several respects. The rocket launch, with its enormous speed and the corresponding pressure on the body, causes stress; weightlessness alters blood circulation and causes the body to age differently. Radiation exposure in space is also increased,” says Andreas Keller, Professor of Clinical Bioinformatics at Saarland University. His research team examined what exactly changes in biological processes in space by analyzing so-called microRNAs—short, non-coding segments of ribonucleic acid (RNA). These regulate the implementation of genetic information within cells. “For this purpose, blood samples from astronauts, such as those taken during earlier space missions for gene analyses as part of NASA’s Twin Study, were not sufficient. Instead, we required tissue samples from mammals,” explains Andreas Keller.

During previous ISS missions, NASA therefore sent several mice into space. These mice were three and eight months old and could be compared with age-matched mice on Earth. The Saarbrücken research team, which worked closely with colleagues at the renowned Stanford University, received 686 small RNA samples from NASA. These samples came from 13 different organs of mice that had spent at least three weeks aboard the International Space Station. “This generated enormous volumes of gene sequencing data, which we analyzed using our bioinformatic methods. These analyses, which built on our many years of experience with microRNAs, took more than a year to complete,” explains Andreas Keller, who also leads a research group at the Helmholtz Institute for Pharmaceutical Research Saarland.

Weightlessness leads to symptoms similar to degenerative diseases

The Saarbrücken scientists focused on how tissue in the heart, brain, spleen, and thymus, as well as in the digestive tract, changes under space conditions. “We found that the physiological effects of spaceflight on humans are considerable. Prolonged stays in weightlessness lead to symptoms similar to degenerative diseases observed on Earth. These include muscle atrophy and bone loss, a weakened cardiovascular system, and changes in the immune system,” explains Andreas Keller. In addition, the team observed that organs age differently in weightlessness, suggesting that astronauts may experience accelerated aging. “These effects intensify with the duration of the mission, which is an important consideration for future missions to Mars and beyond, which would last significantly longer. The goal should now be to identify biomarkers and therapeutic targets through further research in order to mitigate the negative effects on astronauts,” says Andreas Keller.

The research results were published in the renowned scientific journal Nature Communications. The first authors are Friederike Grandke and Shusruto Rishik. The work was conducted under the leadership of Professor Andreas Keller (Saarland University) and Professor Tony Wyss-Coray (Stanford University). A further publication on the identified gene sequencing patterns is planned for the spring.

Original publication:

Friederike Grandke, Shusruto Rishik, Viktoria Wagner, Annika Engel, Nicole Ludwig, Kruti Calcuttawala, Fabian Kern, Verena Keller, Marcin Krawczyk, Louis Stodieck, Virginia Ferguson, Amanda Roberts, Eckart Meese, Nicholas Schaum, Steven Quake, Tony Wyss-Coray & Andreas Keller:

“MiRNAs shape mouse age-independent tissue adaptation to spaceflight via ECM and developmental pathways” in Nature Communications 17, 1387 (2026):

https://doi.org/10.1038/s41467-026-68737-1

Further Information:

https://www.ccb.uni-saarland.de/

https://www.helmholtz-hips.de/de/forschung/people/person/prof-dr-andreas-keller/

Contact for Inquiries:

Prof. Dr. Andreas Keller
Tel. +49 681 302 68611
Mail: andreas.keller(at)ccb.uni-saarland.de

With the 2nd Trusted AI Day, Saarbrücken has once again positioned itself as a central meeting place for discourse on trustworthy artificial intelligence. Representatives from research, business, and politics gathered at the German Research Center for Artificial Intelligence (DFKI) to discuss current developments, key challenges, and concrete solutions in the field of trusted AI.

Congratulations to all our POPL authors!

• • A Verified High-Performance Composable Object Library for Remote Direct Memory Access
    Guillaume Ambal , George Hodgkins, Mark Madler, Gregory Chockler, Brijesh Dongol, Joe Izraelevitz, Azalea Raad, Viktor Vafeiadis

• • Bounded Treewidth, Multiple Context-Free Grammars, and Downward ClosuresRemote
    C. Aiswarya, Pascal Baumann, Prakash Saivasan, Lia Schütze, Georg Zetzsche

• • Endangered by the Language But Saved by the Compiler: Robust Safety via Semantic Back-Translation
    Niklas Mück, Aina Linn Georges,Derek Dreyer, Deepak Garg, Michael Sammler

• • General Decidability Results for Systems with Continuous Counters
    A. R. Balasubramanian, Matthew Hague, Rupak Majumdar, Ramanathan S. Thinniyam, Georg Zetzsche

• • Verifying Almost-Sure Termination for Randomized Distributed Algorithms
    Constantin Enea, Rupak Majumdar, Harshit Jitendra Motwani, V.R. Sathiyanarayana

The ACM is the largest and most important international scientific professional society for computer science. For the year 2026, only 71 new ACM Fellows were selected worldwide; they will be officially honored at an award ceremony on June 13 in San Francisco. The new Fellows were nominated by their colleagues for their outstanding achievements through technical innovation and their contributions to the field. This year’s award recipients come from 14 countries and are among the more than 100,000 members of the ACM worldwide. ACM Fellows serve as ambassadors of the organization and are often asked to make their expertise available to the media, public authorities, and industry leaders.

The research focus of Sven Apel, Professor of Software Engineering at Saarland University, lies in developing methods, tools, and theories with which reliable, efficient, and maintainable software can be constructed and analyzed. For him, the human being as a software developer is at the center, and he also dedicates himself to various interdisciplinary research questions.

Short Biography

Sven Apel studied computer science for a diploma degree at the University of Magdeburg from 1996 to 2002. He completed his doctorate there from 2003 to 2007, also in computer science. His dissertation (awarded summa cum laude) received several prestigious prizes. From 2010 to 2013, he led the Emmy Noether Research Group “Secure and Efficient Software Product Lines” at the University of Passau, before being appointed professor in 2013 within the framework of the DFG Heisenberg Program. Since 2019, Sven Apel has been Professor of Computer Science with a focus on Software Engineering at Saarland University. Since 2022, he has also led the project “Brains on Code,” funded by the European Research Council (ERC Advanced Grant). The project aims to investigate the foundations of program comprehension using neurophysiological methods. Sven Apel is the author or co-author of more than two hundred peer-reviewed publications.

This text was machine translated from the German with no human editing.

Fewer than 0.1% of voting IEEE members receive this prestigious recognition each year. Andreas Zeller is being honored for his outstanding contributions to the analysis of software and its development processes. He has devoted his scientific career to software security: Zeller’s research focuses primarily on automated debugging, mining software repositories, specification mining, and security testing. The long-term international impact of his research is evidenced by a total of nine “Test of Time Awards,” which recognize the enduring influence of individual research contributions. Zeller is also one of the few researchers to have received two ERC Advanced Grants from the European Research Council. Andreas Zeller is an ACM Fellow, a member of the Academia Europaea, and a recipient of both the ACM SIGSOFT Outstanding Research Award and the ACM SIGSOFT Influential Educator Award.

Commenting on his induction as an IEEE Fellow, Andreas Zeller said: “I am very proud that my colleagues have selected me for this special honor from the IEEE. The stated goal of the IEEE is to place technological progress at the service of humanity. In this light, I also see my own work in software security: people need secure applications to participate in the digital world with confidence.”

The IEEE has more than 500,000 members from over 190 countries. It publishes about one third of the scholarly literature in electrical engineering, computer science, and electronics, and supports more than 2,000 conferences annually. The grade of “Fellow” is reserved for experienced IEEE members who have demonstrated an exceptional record of accomplishments in their respective disciplines. Appointment as an IEEE Fellow is preceded by a rigorous evaluation process. Of 1,268 nominations for 2026, only 350 members were ultimately elevated to Fellow status.

This text was machine translated from the German with no human editing.

The world’s leading professional society dedicated to the advancement of artificial intelligence recognizes individuals as Fellows who have typically made substantial and sustained contributions to the field over a period of at least ten years. Ingmar Weber is particularly active in the computational social science community and in the organization of the International AAAI Conference on Web and Social Media (ICWSM). Ingmar Weber is one of only about ten German researchers to date who have been named Fellows of the U.S.-based AAAI.

The achievements of AAAI Fellows range from groundbreaking advances in the theory of artificial intelligence to exceptional accomplishments in AI technology and its applications. The newly elected Fellows have been announced on the AAAI website and will be officially inducted on January 22.

This text was machine translated from the German with no human editing.

Alexander Koller was honored “for his fundamental contributions to computational semantics, grammar formalisms, and neuro-symbolic architectures.” The Fellows Program recognizes ACL members whose contributions to the field are exceptional in terms of scientific and technical excellence, service to the association and the community, or educational and outreach activities.

Further information: ACL Fellows 2025

This text has been machine translated from the German and has undergone no postediting.

Meet Saarland University’s LaLa Lab. Directed by Professor Tomohiro Nagashima, LaLa Lab explores how we learn, how AI can improve learning, and how to make sure learners keep their agency in a society where automated learning is starting to become the status quo. To achieve this, the lab stands at the intersection of human-computer Interaction, the learning sciences, cognitive science, and psychology.

As one of Saarland University’s recently selected Henriette Herz Scouts, Professor Nagashima has been preparing to welcome three international Humboldt Research Fellows. In this interview, he shares his vision for the fellowship, criteria he seeks in applicants, the role of diversity and interdisciplinarity in the field of learning sciences, and the future directions of LaLa Lab.

Could you briefly introduce LaLa Lab and its activities?

“Happy to do so! We’re still a young research group, small but growing at a fast pace. I’m starting my fourth year at the university. LaLa Lab (Learning to Adapt, Learning with Agency Lab) is highly interdisciplinary in its approach and outcomes – that is, both our methodological approaches and publication venues. We use advanced technologies, such as AI-based tutoring systems, to understand how people learn in complex ways. At the same time, we use these technologies to empower learners so that they can learn with agency and adaptively.

Our project AlgeSPACE is designed to help high school students strategically decide which solution approaches they should use for solving systems of equations. At the same time, we collect various types of data to see how exactly human learning happens. This includes student interactions with the system, pre- and post-activity measures to see if students learn by using the system, and qualitative insights such as interviews and think-aloud responses to deeply investigate their experience with the system. 

A screenshot from AlgeSPACE.

For example, in our Mindful Math project, we also measure how a mindful, LLM-based conversation experience with our virtual fox agent can help students reduce their math anxiety while supporting their learning. 

A screenshot of the Mindful Math tutor.

To conduct research that is meaningful for both researchers and practitioners, we work very closely with school students, teachers, administrators, and parents in Saarland and beyond, also internationally in North America and Asia. Together, we co-design technologies that serve not only our research purposes but also teachers and students in real classrooms. Technologies like AlgeSPACE are freely and openly available, and have been used in everyday teaching activities.”

Could you share a bit about your vision or plans for bringing fellows into LaLa Lab over the next couple of years through this Humboldt Research Fellowship?

“When I applied for this opportunity to become a Henriette Herz Scout, I was hoping to make our lab even more interdisciplinary. We are already quite diverse — with students from computer science, educational technology, psychology, and even cultural studies — but I wanted to boost this strength further.

I plan to bring in three postdocs in the next few years, each with a different expertise. Since I work at the intersection of the learning sciences, human-computer interaction, and psychology/cognitive science, my idea is to select fellows from each of these disciplines to strengthen our approach, so they can also learn from each other and gain new skills from other disciplines in the lab.” 

What do you think separates the Humboldt Research Fellows, recruited through Henriette Herz Scouts, from other postdoc opportunities for international postdocs?

“One thing that I am committed to is giving the fellows a sense of what a PI’s job looks like. A particularly generous aspect of this program is that they get their own research funding of approximately 20,000 euros. My plan is to let them use this funding to simulate the experience of having a ‘mini’ lab within our lab.

They can decide what they want to do: have their own projects, decide who to hire, do recruiting and interviewing, figure out how to allocate their resources to run a project effectively… All this decision-making is something that regular postdocs wouldn’t necessarily get to experience. That’s one of the biggest opportunities I see. After two years, I hope that the three fellows will feel ready to lead their own labs with the skills they gained.” 

How do you see the fellowship support the career paths of the fellows, in addition to their research? 

“The fellows will be experiencing the PI role in a risk-free environment, because I’m still responsible for all the decision-making in the end. Personally, I jumped from finishing my PhD to a professor position. There are a lot of things you learn as you go — and I think I would’ve liked to get management and leadership experience in a risk-free environment as a postdoc. This opportunity will familiarize them with the different types of decisions a PI has to make, supporting a smoother transition when they become a PI.” 

While you’re reviewing applications, what are some qualities you pay attention to in a candidate?

“There are several aspects. One thing that I want to ensure — which is also aligned with the fellowship — is bringing international talent to Germany, especially those who might not otherwise have had the opportunity to come here. I tried to share this job opportunity through various channels — mailing lists, social media, conferences, academic societies, and personal connections — to reach talented candidates around the world.

Our lab is already quite diverse, but as I mentioned, we want to boost this strength even further. I’m happy to see that I’m getting applications from all over the world. I’m very excited about the different perspectives they will bring — they’re trained in different education systems and have diverse cultural views, which are important for my field of study and for running a diverse and inclusive lab. 

In addition, to further expand our interdisciplinarity, I didn’t focus on just one expertise when I made a job ad for the first Humboldt Fellow position. I decided to target broader fields, so applicants could bring expertise in designing systems, understanding human behaviors, analyzing complex data, and/or technical implementations. I wanted to see what they might be able to bring in, and that’s something you can’t really know until you get the applications. That’s why I’m interviewing potential candidates from different fields — to decide who to accept in the first round and who to accept next — so they effectively complement each other in terms of disciplinary and interdisciplinary skills when they join.

Last but not least, I’ve also emphasized leadership and management potential during the interviews and in the job listing, which candidates particularly liked. I want to see if they can be independent in coming up with research ideas, doing management tasks or mentoring students. To assess this, I asked applicants for a cover letter outlining an independent project idea that they would like to lead and manage in the lab.”

Do you have any criteria other than publications/academic record that could mark someone as a strong candidate? Any particular skills or mindsets that you might find specifically valuable in a postdoc working in LaLa Lab?

“Although these decisions can be difficult to articulate, something I always ask in interviews is candidates’ attitudes toward collaboration with people from different backgrounds. Since our lab collaborates across fields, it’s often necessary for candidates to be able to explain things from their background to someone who isn’t familiar with that concept. They need to communicate their work to people from design and technical backgrounds, as well as stakeholders in the community who may not know a lot about research. Doing community-based research in my field requires actively listening to others’ thoughts, effectively negotiating with people from different backgrounds and daily experiences, and finding solutions that satisfy all parties involved – this is very hard! 

In addition to this, I pay attention to their ability to mentor students and scaffold their learning, since supervision skills require time to develop and people have different mentoring styles. I want to know how they approach mentoring.” 

What kind of independence and freedom will fellows have when it comes to proposing and leading their own projects?

“In my lab, fellows can propose and work on their own projects based on their interests and research, as long as the project fits within the growth scope of the lab. The fellowship funding can be used for these independent projects, and I will be happy to help or compensate for anything the funding doesn’t cover, if needed. In addition to their own project,  I will also find an existing project where they can be integrated.

I think this is a nice opportunity; they can actually lead something independently while also becoming well integrated within the lab by getting involved in existing projects. This way, they learn how things are done here and meet the other lab members, while I can still provide guidance as they work on their own.”

What excites you most about welcoming new international researchers into LaLa Lab?

“Getting international people itself is very exciting. I come from Japan, but my training was in the USA, and now in Germany I see many advantages that people bring through their diverse training and cultural experiences. 

For example, the way we conduct design research with schools here is different from what I was used to in the USA. There are always new tactics and approaches to learn based on the cultural setting. Technologies are used in different ways across cultures. Teachers here have a different approach to research than their counterparts in the USA. These experiences give you a new perspective on how to adjust your approach to what’s in front of you.

At the same time, I want to learn from their experiences in other countries I’m less familiar with to see what other strategies we might be able to integrate in the lab. I think this will help the lab become more flexible and improve our approaches to conducting our work.”

What advice would you give to international researchers who might consider applying in the future?

“The hiring process is quite flexible. I can hire three postdocs over the next few years, with possible overlaps and different start dates. That’s why I’m planning to have another call sometime soon to bring in another person. 

One thing I can say is that our lab is growing not only in size but also in approach. As a lab, we’re trying to adapt our strategies to make sure we move in the right direction and build on existing work. So, future candidates may discover opportunities that are more relevant to them if they keep up with our work.

So, my advice: check my website and stay updated!”

What kind of projects and directions are you hoping to head in Lala Lab in the next few years?

“That’s a big question. In the lab, our work is organized around several pillars, and each of them has multiple sub-projects. For example, something we’re increasingly focusing on is how complex dynamic interactions happen between teachers, students, and AI systems used in classroom settings. There’s a lot of dynamic decision-making involved in the process.

This decision-making is hard to understand, but also very interesting to explore. One thing that we’re prioritizing is ensuring that the human stakeholders, like teachers and students, keep their agency while making decisions, since technologies with ‘accurate’ models may attempt to detect the students’ emotional or cognitive state and make decisions for them (more AI control).

If that happens too much, sometimes it can take away important decision-making opportunities for students. Students might want to decide what topics to learn by themselves, instead of getting assigned one based on their previous performance. These types of decision-making opportunities help them become more metacognitive and self-regulated.

So that’s something we will discuss more in the future: how we can ensure that students and teachers control their own decision-making, and how to balance that control between AI systems and human stakeholders to use AI systems not only effectively but also ethically and sustainably.”

Editor’s note: The interviewee wishes to acknowledge the contributions of lab members Katharina Bonaventura, Mirella Hladký, Helene Nüttgens, Vera Rief, Mareike Silber, and Man “Echo” Su,  to the projects referenced.

Want to find out more about the research conducted in LaLa Lab? Visit their website.

Are you interested in creating ethical AI systems that support students? Are you an international researcher? Maybe you could be the next Henriette Herz fellow in one of Germany’s most interdisciplinary learning-science labs. Take a look at the application requirements.

Relevant pages:

Computer scientists at Saarland University develop customized learning software

Information about the Humboldt Research Fellowship

Editor:
Saarland Informatics Campus Team
Email: yagmur.akarsu@uni-saarland.de

Realistic digital avatars are becoming increasingly relevant, for example in virtual and augmented reality applications, video conferencing, films and computer games, or in medicine. Researchers at the Max Planck Institute (MPI) for Informatics in Saarbrücken, Germany, are now presenting two novel methods at two of the world’s leading computer graphics conferences, SIGGRAPH and SIGGRAPH Asia. These methods enable the generation of photorealistic full-body avatars and allow head avatars to be controlled using only audio tracks.

Previous methods for generating digital avatars have had significant limitations: The face and body often cannot be controlled independently, clothing sometimes looks unnatural, the renderings are often convincing only from certain perspectives, and facial animations frequently appear sterile and lifeless. With their works “EVA: Expressive Virtual Avatars from Multi-view Videos” and “Audio-Driven Universal Gaussian Head Avatars”, the Max Planck researchers are taking a step toward solving these problems.

The paper “Audio-Driven Universal Gaussian Head Avatars”, to be presented in December at SIGGRAPH Asia in Hong Kong, describes a method by which photorealistic 3D head avatars can be automatically animated and controlled using only voice recordings. The foundation of this is the newly developed Universal Head Avatar Prior (UHAP), a model pre-trained on a large number of video recordings of real people from a publicly available dataset. It can clearly distinguish between identity (the appearance of a specific person) and expression (facial expressions and movements).

Marc Habermann, Kartik Teotia and Christian Theobalt (left to right). Photo: MPI-INF/Philipp Zapf-Schramm

An audio-encoder then translates audio signals directly into the expression renderings of the digital avatar model. Unlike earlier approaches, it takes into account not only lip and jaw movements but also fine, audio-dependent changes such as movements inside the mouth or subtle facial expressions. Using this pre-trained model, highly realistic 3D facial renderings can be generated with significantly less data. “Our goal is to create digital heads that not only synchronize with speech, but also behave lifelike, incorporating subtle details such as eyebrow movements and gaze shifts,” says Kartik Teotia, a doctoral student at Saarland University conducting research at MPI for Informatics.

In addition to faces, the research at MPI for Informatics also covers methods for generating full-body avatars. The paper “EVA: Expressive Virtual Avatars from Multi-view Videos”, published in August at the SIGGRAPH conference in Vancouver, describes a novel approach in which the modeling of motion and appearance are separated. A flexible digital model first captures the body, hands, and face, along with their movements and expressions. A second layer then adds the external appearance, that is, skin, hair, and clothing. “With EVA, we can realistically generate movements and facial expressions independently of one another, and also render them from new viewpoints that were not inlcuded in the original recordings,” says Marc Habermann, head of the research group Graphics and Vision for Digital Humans at MPI for Informatics. For now, the system still requires training with recordings from a lab facility at the Institute, where a person is filmed from more than one hundred camera perspectives simultaneously.

“With these two works, we are advancing research on realistic digital avatars in a decisive way. Such models could fundamentally change how we communicate, collaborate, or acquire new skills in the future, for example through virtual tutors, extending far beyond computer science,” says Professor Christian Theobalt, Director at the Max Planck Institute for Informatics and head of the Visual Computing and Artificial Intelligence department, where these projects are being developed. Theobalt is also founding director of the Saarbrücken Research Center for Visual Computing, Interaction and Artificial Intelligence (VIA), a strategic research partnership with Google.

Both of the above-mentioned works have already attracted interest from industry. “EVA: Expressive Virtual Avatars from Multi-view Videos” was developed in collaboration with Google at the Saarbrücken VIA Center. “Audio-Driven Universal Gaussian Head Avatars” was developed with scientific collaboration from Flawless AI, a London-based film technology company recently named one of TIME Magazine’s 100 Most Influential Companies of 2025. Flawless AI’s Visual Dubbing technology, built on foundational research pioneered by Theobalt’s department, enables actors’ lip movements to be precisely adapted for new languages, a breakthrough that is drawing growing attention across Hollywood. In May 2025, the first full-length feature reworked with Visual Dubbing, Watch the Skies, was released in U.S. cinemas.

Original publications:
Kartik Teotia, Helge Rhodin, Mohit Mendiratta, Hyeongwoo Kim, Marc Habermann, and Christian Theobalt. 2025. Audio-Driven Universal Gaussian Head Avatars. In SIGGRAPH Asia 2025 Conference Papers December 15–18, 2025, Hong Kong, Hong Kong. ACM, New York,NY, USA, 16 pages. https://doi.org/10.48550/arXiv.2509.18924

Hendrik Junkawitsch, Guoxing Sun, Heming Zhu, Christian Theobalt, and Marc Habermann. 2025. EVA: Expressive Virtual Avatars from Multi-view Videos.In Special Interest Group on Computer Graphics and Interactive TechniquesConference Conference Papers (SIGGRAPH Conference Papers ’25), August10–14, 2025, Vancouver, BC, Canada. ACM, New York, NY, USA, 20 pages. https://doi.org/10.1145/3721238.3730677

Further Information:
– Website of the Visual Computing and Artificicial Intelligence department: https://www.mpi-inf.mpg.de/de/departments/visual-computing-and-artificial-intelligence
– Website of the Graphics and Vision for Digital Humans group: https://gvdh.mpi-inf.mpg.de/
– Saarbrücken Research Center for Visual Computing, Interaction and Artificial Intelligence: https://www.via-center.science/

Scientific contacts:
Prof. Dr. Christian Theobalt
Director, Department „Visual Computing and Artificial Intelligence“
Max Planck Institute for Informatics
Mail: d6-sek@mpi-inf.mpg.de
Tel: +49 681 9325 4500

Dr. Marc Habermann
Group leader, Group „Graphics and Vision for Digital Humans“
Max Planck Institute for Informatics
Mail: mhaberma@mpi-inf.mpg.de
Tel: +49 681 9325 4507

Press contact and editor:
Philipp Zapf-Schramm
Max Planck Institute for Informatics
Phone: +49 681 9325 4509
Email: pzs@mpi-inf.mpg.de

On the evening of December 4, 2025, around 80 students and several professors gathered in front of building C7.4 and walked with torches across campus to the computer science department and back, before rounding off the event with hot mulled wine and non-alcoholic punch in C7.2.

“I am extremely grateful to the students for this remarkable sign of appreciation,” says Alexander Koller. “Torchlight processions are a beautiful tradition in Saarbrücken computer science, and I feel deeply honored that so many students have taken up this tradition in our field as well.”

Alexander Koller is Professor of Computational Linguistics in the Department of Language Science and Technology and also holds a courtesy professorship in Computer Science, both at Saarland University. He heads the Computational Linguistics group and serves as speaker of the DFG-funded Research Training Group “Neuroexplicit Models of Language, Vision, and Action”. His research focuses on computational models of meaning and reasoning in natural language processing, particularly dialog systems, interactive natural language generation and neurosymbolic approaches to semantic parsing. His work has received multiple distinctions at leading conferences such as the Annual Meeting of the Association for Computational Linguistics (ACL), and he is a member of the ELLIS Saarbrücken unit.

Background Saarland Informatics Campus:
900 scientists (including 400 PhD students) and about 2,500 students from more than 80 nations make the Saarland Informatics Campus (SIC) one of the leading locations for computer science in Europe. Four world-renowned research institutes cover the entire spectrum of computer science, namely the German Research Center for Artificial Intelligence (DFKI), the Max Planck Institute for Informatics, the Max Planck Institute for Software Systems, the Center for Bioinformatics as well as Saarland University with three departments and 24 degree programs.

Researchers from Saarland University and the Max Planck Institute for Software Systems have, for the first time, shown that the reactions of humans and large language models (LLMs) to complex or misleading program code significantly align, by comparing brain activity of study participants with model uncertainty. Building on this, the team developed a data-driven method to automatically detect such confusing areas in code — a promising step toward better AI assistants for software development.

At the invitation of the German Research Center for Artificial Intelligence (DFKI), scientists from the university centres BIFOLD, Lamarr Institute, MCML, ScaDS.AI Dresden/Leipzig, and Tübingen AI Center shared and discussed their latest research findings in Saarland University’s assembly hall. This year, the French AI community also participated for the first time, represented by the AI competence centres 3IA Côte d’Azur, ANITI, DATAIA, ENACT, PostGen@Paris, PR[AI]RIE-PSAI, SEQUOIA, and the national computer science institute Inria.

The German AI competence centers – five of which are located at universities – and DFKI form a national network that serves as the foundation of the German scientific AI ecosystem. As part of the German government’s AI strategy, the five university AI competence centres have been permanently funded by the federal government and the respective states since July 1, 2022. The federal government allocates up to €50 million annually for this initiative. DFKI receives up to €11 million per year from the Federal Ministry of Research, Technology, and Space (BMFTR) as part of its upcoming project funding program.

In his welcoming speech, Matthias Hauer, Parliamentary State Secretary at the Federal Ministry of Research, Technology and Space, emphasized: “Research will help determine what the AI of the future and a future with AI will look like. The German government wants AI that is ‘Made in Germany’: secure and trustworthy, resource-efficient and therefore sustainable, oriented toward the common good and serving people, not replacing them. Germany is one of the world’s leading research locations. The AI competence centers stand for excellent research and strengthen Germany’s position as a top AI location. In order to achieve the goals of the High-Tech Agenda Germany, we now want to transfer our excellent AI research to the economy – for greater competitiveness, added value, and sovereignty.”

The meeting between the German and French network partners is the result of DFKI’s longstanding relationship with Inria. Since 2018, Inria has been coordinating the French government’s national AI strategy. Both organizations have collaborated on projects for decades, initially at the Saarbrücken and Nancy locations. On January 22, 2020, the institute-wide cooperation was officially formalized.

Antonio Krüger, who will lead the scientific committee of the AI competence centres in 2025, emphasized the importance of cross-border collaboration: “I am particularly pleased to welcome the French national computer science institute Inria and the French AI clusters to the All Hands Meeting of the AI Competence Centres Network. The design of AI is increasingly becoming a central operating system for the economy and society. Designing it in such a way that we can rely on it as we do on electronic calculators is a major challenge, but also a great opportunity for Germany, France, and the European Union. We have the ecosystems and, with the AI Gigafactories, the computing infrastructure to meet these challenges.”

Pierre Alliez, Senior Team Leader at Inria Sophia-Antipolis and Scientific coordinator of the partnership between Inria and DFKI, highlighted the significance of a joint Franco-German AI strategy: “Today’s event is the first scientific pilot meeting of the French AI cluster network outside France. This event is therefore also a milestone in the Franco-German joint agenda in the field of AI and marks the start of a structured initiative between French and German players.”

Jürgen Barke, Saarland’s Minister for Economic Affairs, Innovation, Digital, and Energy, emphasized the vital role of artificial intelligence in shaping the transformation process and the future of work and production. In Saarland, AI is a key driver of economic change and competitiveness.

JJérôme Spinoza, the French Consul General in Germany, stated that the meeting captured the core of the Franco-German partnership in artificial intelligence and beyond. The focus was on the shared goal of advancing research and innovation to foster European integration and boost Europe’s momentum.

On November 5, the meeting mainly centered on intensive scientific exchange and networking between the German and French AI centres.

AI gigafactories for European sovereignty in the field of artificial intelligence

November 6 provided a look into Germany’s AI infrastructure, the efforts of AI service centres, and the bright future for young AI researchers, whether in academia or starting new companies.
Dr. Ferri Abolhassan, a member of the Board of Management at Deutsche Telekom AG and Chairman of the DFKI Supervisory Board, sparked a panel discussion on AI gigafactories, emphasizing the challenges and opportunities for German and European AI sovereignty. AI gigafactories with 100,000 GPUs are essential for developing and deploying AI in science and industry. Moderated by DFKI corporate spokesperson Reinhard Karger, the panel included Mirko Holzer (Federal Agency for Breakthrough Innovations – SPRIND), Prof. Dr. Holger Hoos (Alexander von Humboldt Professor for Artificial Intelligence, RWTH Aachen University), and Prof. Dr. Feiyu Xu (German University of Digital Science, member of the supervisory boards at Airbus SE, Chain IQ Group AG, ZF Friedrichshafen, and Siemens Energy AG). They discussed AI lighthouse applications in healthcare, such as personalized drug development and improved hospital management; breakthrough innovations in startups rather than incremental improvements; industrial AI through automated factories; AI to reduce bureaucracy; and hybrid defense using AI for cybersecurity. The consensus was clear: Europe must and can develop AI innovations independently. The six AI competence centres are encouraged to pursue the opportunities presented by the EU’s AI Continent Action Plan, the AI for Science initiative, and the Apply AI strategy at the European AI in Science Summit in Copenhagen on November 3-4.

Academic career or startup?

What specific challenges do founders of technology- and research-based startups face? How can aspiring entrepreneurs gain from networking opportunities like Grid AI, and what obstacles must be overcome when transitioning from research to startup? What advantages do startup and innovation platforms provide for entering the French and German markets – especially for startups with a European focus?

Moderated by DFKI co-speaker Andreas Schepers, the panel discussion featured Matthias Schmitz (Southwest X), Charlotte Peyrat-Vaganay (Inria), and Laure Poirson (AI Grid) and focused on career paths for computer science and AI graduates and on opportunities for successfully translating research results into entrepreneurial practice.

Conclusion and outlook
The meeting impressively demonstrated how German AI competence centres are joining forces and paving the way for a networked European AI landscape together with the French research community.

Further information
https://www.bmftr.bund.de/EN/Research/EmergingTechnologies/ArtificialIntelligence/artificialintelligence_node.html

All Hands Meeting 2025 Website: https://all-hands-meeting-2025.dfki.de

Press Contact:
Heike Leonhard, M.A.
Communications & Media DFKI Saarbrücken
Heike.Leonhard@dfki.de
Tel.: +49 681 85775 5390

DFKI press release: https://www.dfki.de/en/web/news/german-and-french-ai-competence-centres-strengthen-european-ai-sovereignty

DFKI press release: https://www.dfki.de/en/web/news/dfki-supervisory-board-new-chair

AI assistants have rapidly become part of everyday life – whether through our interactions with large language models like ChatGPT or in medical applications that help interpret complex datasets. Yet, despite their widespread use, AI systems still make surprisingly simple mistakes that persist even after extensive training. They also lack the ability to think logically or to truly ‘understand’ nested input. For Michael Hahn the problem lies in the architecture itself.

Outstanding research performance in academia is often measured by how frequently a scholar’s work is cited by others. Each year, the ‘Highly Cited Researchers’ list recognizes leading experts in specific fields whose publications are among the most frequently cited worldwide. This year, three researchers from Saarland University have once again earned a place on the prestigious list, reflecting their exceptional academic reputation.

The fellowship is in the area of Algorithms and Optimization.

Yonggang Jiang is now supported by the Google PhD Fellowship. This program recognizes and supports outstanding doctoral candidates conducting innovative research in computer science and related fields. The financial support can run for up to two years; a reviewal will be conducted after one year.

Yonggang is particularly interested in the design and analysis of algorithms. Even since his Highschool time he was more interested in topics in combinatorics than calculus or the like. He says that he finds beauty in discreet mathematics.

 

Shortcutting graphs makes algorithms perform much better. Picture: Yonggang Jiang

 

In 2021 he graduated with a Bachelor’s degree from the Computer Science and Technology Department of Nanjing University. In 2023 he joined the Max Planck Institute for Informatics as a PhD student in the department of Danupon Na Nongkai.

His current research focuses on graph algorithms, with the goal of understanding their theoretical limits across various settings, including parallel and distributed computing. His vision is to deepen the understanding of graph theory and algorithms.

Further Information:

https://yonggangjiang.github.io/

Editor:
Bertram Somieski
Joint Administration
Max Planck Institute for Informatics
Max Planck Institute for Software Systems
Tel: +49 681 9302-5710
E-Mail: somieski@mpi-klsb.mpg.de

The Ackermann Award is an international prize presented annually to the author of an exceptional doctoral dissertation in the field of Computer Science Logic.