Header logo is de

Unabhängige Forschungsgruppen

Autonomes Maschinelles Sehen

Max-Planck-Forschungsgruppe für Autonomes Maschinelles Sehen

Intelligente Autonome Systeme müssen ihre Umgebung effizient und robust wahrnehmen um sich in ihrer komplexen, veränderlichen Welt zurechtfinden zu können. Eine Schwierigkeit dabei besteht in der Umwandlung der gewaltigen Menge an eingehender, mehrdeutiger und unvollständiger Information in eine einfache und kompakte Repräsentation. Zur dreidimensionalen Interpretation der Szene muss zudem die durch den Projektionsprozess verlorene Information wieder hergestellt werden.

Autonomous Learning

Dynamische Lokomotion

Die Max-Planck-Forschungsgruppe Dynamische Lokomotion wird von Dr. Alexander Spröwitz geleitet. Robotic Biomechanics ist sein Spezialgebiet, denn er sieht sich als Robotiker und Biomechaniker. Tiere laufen dynamisch und effizient, elegant und adaptiv. Ihre Fortbewegung ist ein sorgfältig orchestriertes Zusammenspiel der Muskeln und Sehnen, das im Laufe der Evolution immer mehr optimiert wurde. Spröwitz und sein vierköpfiges Doktoranten-Team nutzen Roboter, um Tiere und deren Bewegungsabläufe zu verstehen. Wie aktiviert ein Tier einen Muskel? Welche Kräfte sorgen dafür, dass sich das Tier fortbewegen kann? Diese Vorlagen aus der Biomechanik verwenden die Forscher, um daraus Roboter-Modelle zu bauen. Die vielen Einzelteile, die es beim Tier gibt (Muskeln, Sehnen und Knochen) verwenden sie, um ein vereinfachtes Modell eines Tieres zu b... Read More

Max Planck Independent Research Group for Embodied Vision

We research fundamentals of intelligent embodied agents such as robots that learn to perceive and act through interaction with their environment. Our group investigates novel methods for learning the basic physical 3D understanding of dynamic environments up to complex tasks such as autonomous navigation and object manipulation from raw sensory measurements and environment interactions. Besides vision as a primary sensing modality, we also consider further sensing modalities such as tactile or proprioceptive sensing.

Intelligent Control Systems

Research in the Intelligent Control Systems group focuses on decision making, control, and learning for autonomous intelligent systems. We seek to develop fundamental methods and algorithms that enable robots and other intelligent systems to interact with their environment through feedback, autonomously learn from data, and interconnect with each other to form collaborative networks. Turning mathematical and theoretical insight into enhanced autonomy and performance of real-world physical systems is an important and driving facet of our work.

Locomotion in Biorobotic and Somatic Systems

We study how animals move using non-invasive, state of the art, motion observation techniques (e.g. high speed video, pose estimation based on deep learning) both in their natural habitat and in the lab in cooperation with academic partners around the world. To better understand biological movement as observed first hand in natural systems through original discovery, we also build biorobotic models that emulate animal locomotion. Bodies are in nature are soft, and so our approach leverages advances in Soft Robotics. In this spirit, we make diverse soft ‘artificial muscles’ (actuators) and soft sensors. Discovery of principles of animal locomotion can then be transferred to biologically informed robots where advantageous. These robophysical models therefore serve as instruments of knowledge. Robotics-inspired Biology can help us better understand the spectacular and graceful movement observed in natural systems.

Locomotion in Biorobotic & Somatic Systems is an independent Max Planck Research Group at the interdisciplinary Max Planck Institute for Intelligent Systems (formerly Metals and Materials Science). Leveraging expertise in diverse fields as biology and robotics, we explore the behavioural and morphological adaptations of natural systems, to achieve robust multimodal locomotion. Richard Feynman famously said: “What I cannot create, I do not understand”. That is, if we cannot recreate how animals move, we have not understood precisely how it works. Curiosity driven fundamental research is therefore required to decipher how animals are able to gracefully navigate irregular terrain with such agility and robustness. The past decade has presented a dramatic expansion in the development of mobile robots and the application of robotic systems to practical tasks. Despite the proliferation of computation and sensing at the small scale, robots still remain largely unable to a... Read More

Micro, Nano, and Molecular Systems

Our group has broad interests in the interaction of optical, electric, and magnetic fields with matter at small length scales. We work on new 3-D fabrication methods, self-assembly, actuation, and propulsion. We have observed a number of fundamental effects and are developing new experimental techniques and instruments.

Movement Generation and Control

Our research applies in a broad range of fields, from medicine and psychiatry to social and communication systems. Recently, we also began putting a special focus on consequential decision making in several domains, including hiring processes, pre-trial bail, or loan approval.

The Research Group on Probabilistic Learning focuses on improving the flexibility, robustness and ethics of machine learning methods for real-world applications. Flexible means they are capable of modeling complex real-world data, which are often heterogeneous in nature and present temporal dependencies. Secondly, we aim to improve their robustness to outliers, missing data and mixed statistical data types. Finally, we work on aligning algorithms with the ethics of society by making them fairer and interpretable – if algorithms are part of important decision-making processes, the outcomes should be fair and explainable.

Probabilistische Numerik

Max-Planck-Forschungsgruppe für Probabilistische Numerik

Die Forschungsgruppe studiert und entwickelt numerische Verfahren, insbesondere zur Verwendung in intelligenten Systemen. Kern unserer Arbeit ist die Beobachtung, dass Algorithmen zur Berechnung von nicht-analytischen Größen, wie Integralen und Extremwerten, als Inferenz, als lernende Maschinen beschrieben werden können. Es ist daher möglich, klassische numerische Verfahren so zu adaptieren, dass sie hilfreiche Struktur und erschwerende Unsicherheitsquellen, welche gerade in intelligenten Systemen auftreten, explizit modellieren und nutzen, um robustere oder effizientere Antworten zu liefern.

We develop a scientific foundation and practical tools for empowering people to choose and successfully pursue their ideal self and to make valuable contributions to society.

Welcome to the website of the Max Planck Research Group for Rationality Enhancement! We develop a scientific foundation and practical tools for empowering people to choose and successfully pursue their ideal selves and to make valuable contributions to society. Our international, interdisciplinary team combines methods from computational cognitive science, psychology, human-computer interaction, and artificial intelligence. Check out our Facebook page. Follow our lab on twitter Read More

Statistical Learning Theory

Intelligente Nanoplasmonik

Auf der Nanoskala treffen sich die verschiedenen Disziplinen Chemie, Biologie und Materialwissenschaften. Das Feld der Nanoplasmonik befaßt sich dabei speziell mit der Konzentration und Manipulation von Licht in einem Volumen, das im Durchmesser nur wenige Nanometer groß ist. Die Schlüsselmaterialien zur Verwirklichung von nanoplasmonischen Systemen sind Metalle. Die optischen Eigenschaften von Metallen haben die Menschen schon seit Jahrhunderten fasziniert. Wenn Licht auf ein Metall Nanopartikel trifft (z.B. ein kleines Goldteilchen in einem Kirchenfenster) werden kollektive Schwingungen von Metallelektronen angeregt, die man Partikel-Plasmonen nennt.