Max Planck Stanford Center To Be Established

Max Planck Institute for Informatics and Stanford School of Engineering to establish a joint AI research centre

• Funding: The Max Planck Center for Visual and Multimodal Artificial Intelligence is receiving around 29 million euros, of which 11.6 million euros comes from the Saarland, and is collaborating with industry partners such as Google and Flawless. Its main base is the Max Planck Institute for Informatics in Saarbrücken.
• Care robot applications: The Center aims to develop AI systems that process information multimodally, for example for care robots that can understand speech, gestures and visual signals and respond realistically.
• Scientific exchange: The Center is headed by leading scientists from Saarbrücken and Stanford and offers researchers opportunities to spend time at the Stanford School of Engineering.

Advancing future technologies and developing the foundations for a new generation of artificial intelligence systems that can learn from people, interact with them seamlessly, and assist them - this is the aim of the new Max Planck Center for Visual and Multimodal Artificial Intelligence (MPC-VMI). The Center brings together the Max Planck Institute (MPI) for Informatics, Stanford Engineering, and select research partners from Industry in a collaborative research initiative. The Center is financed with around 29 million euros, with the State of Saarland supporting the establishment of the Center with 11.6 Million Euros from the Transformation Fund.

Partners from industry

The decision to establish the new Center followed a highly competitive selection process in the Max Planck Society in which the proposal prevailed over numerous proposals from other locations. The Center will be led by directors Christian Theobalt (MPI for Informatics) and Gordon Wetzstein (Stanford Engineering), together with co-directors Bernt Schiele (MPI for Informatics) and Karen Liu (Stanford Engineering), and the scientific coordinator Marc Habermann (MPI for Informatics). Additionally, the Center is supported by Google and Flawless as two industry research partners.The Center is currently in its founding phase and is expected to begin operations at the end of 2026.

Researchers at the Center will have the opportunity to undertake research stays of up to two years at Stanford School of Engineering. At full capacity, up to 50 researchers are expected to conduct research at the Center. The Center and its partners will benefit both from Stanford's outstanding research environment and its proximity to Silicon Valley, as well as from its location on the Saarland Informatics Campus (SIC), home to numerous internationally renowned computer science research institutes. The Center opens up new possibilities for collaboration with SIC institutions and the Saarland research ecosystem as a whole.

Perspective multimodal AI

Great expectations are being placed on the next generation of AI systems, whether as robotic caregivers or personalized tutors, through new forms of telecommunication, or in human-robot collaboration in manufacturing. To enable such systems to support humans effectively and operate safely and reliably in the complexity of the real world, a number of fundamental research questions must first be addressed, including challenges in the fields of computer vision, computer graphics, machine learning, natural language processing, human-computer interaction, and robotics. The Max Planck Center for Visual and Multimodal Artificial Intelligence aims to play a leading role in the development of such systems.

The example of the care robot illustrates the wide range of scientific challenges involved. As a multimodal and physical AI system, they must be capable of processing various types of information. They require advanced capabilites in computer graphics and computer vision to model and understand their surroundings using novel world models. This requires innovative AI models that can perceive and model complex human interactions, even when only minimal sensor data is available. They must be able to generate, interpret, and respond to all forms of human communication, including spoken language, written text, and body language. They must be able to process visual information reliably in order to navigate safely through a wide variety of environments. In addition, their user interfaces must be intuitive, with elements such as digital faces rendered at the highest levels of realism and detail.

"With the MPC-VMI, we are bringing together in Saarland a critical mass of innovation capacity from world-leading partners in research and industry to advance transformative developments such as these," says Christian Theobalt, Scientific Director at the Max Planck Institute for Informatics and one of the directors of the new Max Planck Center.

Saarland Minister of Finance and Science Jakob von Weizsäcker adds: "With Stanford Engineering and internationally renowned industry partners such as Google's VIA Center and Flawless AI, this new Center brings together an exceptional network of expertise in artificial intelligence. The Center bolsters the international standing and network of Saarland's research ecosystem in computer science even further. With the support of the Saarland Transformation Fund, we are in a position to enable such world-class partnerships."

Max Planck Centers are a cornerstone of the Max Planck Society's internationalization strategy. Within these Centers, researchers collaborate with leading international partners to leverage each other's expertise and resources and advance knowledge in future-oriented fields of research. The new MPC-VMI will be headquartered at the Max Planck Institute for Informatics in Saarbrücken.