AI Colloquium
The AI Colloquium is a series of lectures dedicated to cutting-edge research in the field of machine learning and artificial intelligence, coorganized by the Lamarr Institute for Machine Learning and Artificial Intelligence (Lamarr Institute), the Research Center Trustworthy Data Science and Security (RC Trust), and the Center for Data Science & Simulation at TU Dortmund University (DoDas).
Programme
Distinguished researchers deliver captivating lectures followed by vibrant discussions. However, unlike traditional colloquia, the AI Colloquium prioritizes interactive dialogue, fostering international collaboration. Conducted primarily in English, these 90-minute sessions feature hour-long lectures and 30-minute Q&A sessions. Join every Thursday at 10 AM c.t. for a stimulating exploration of cutting-edge topics. Whether in-person at our Lecture Room on Fraunhofer Strasse 25 or via Zoom, our hybrid format ensures accessibility for all.
| Day (usually) | Thursday |
| Start and end time | 10 AM c.t. - 12 AM |
| Duration of Presentation | 60 Minutes |
| Location (usually) | Lecture Room 303 3. Floor Fraunhofer Strasse 25 Dortmund |
Upcomming Events
Off the Grid - Neural Representations and Neural Operators for Image Processing
- Lamarr
Abstract: While images are classically represented on a regular grid of pixels, recent research has become increasingly interested in neural representations, i.e., the representation of scenes via parameterized continuous functions similar to neural networks. In this talk I will discuss particular aspects of such representations: I will show how neural representations of segmentation masks allow enforcing constraints that are difficult to enforce in a classical pixel-wise segmentation, and illustrate how suitable neural representations of a video allow decomposing scenes in such a way that individual objects can be edited easily. Finally, I will discuss that the perspective of viewing images as functions leads to common image-to-image networks becoming neural operators, and illustrate different architectures for such neural operators at the example of improving the reconstruction of linear inverse problems.
Prof. Michael Möller
Bio: Michael Möller studied and completed his doctorate in applied mathematics in applied mathematics in Münster from 2004 to 2012, during which time he spent two years as a researcher at the University of California, Los Angeles, UCLA during this time. After completing his doctorate with his supervisor Martin Burger, he worked for 1.5 years in the research and development department department of Arnold & Richter Cine Technik GmbH (ARRI) before joining the computer vision group of Daniel Cremers at the Technical Cremers’ computer vision group at the Technical University of Munich. Since 2016 he is a professor at the University of Siegen. His main field of research is the combination of model- and learning-based methods for image reconstruction and analysis. He is part of the DFG priority program “Theoretical Foundations of Deep Learning”, spokesperson of the DFG research group “Learning to Sense” and was awarded a Lamarr Fellow in 2023.
Past Events
Off the Grid - Neural Representations and Neural Operators for Image Processing
- Lamarr
Abstract: While images are classically represented on a regular grid of pixels, recent research has become increasingly interested in neural representations, i.e., the representation of scenes via parameterized continuous functions similar to neural networks. In this talk I will discuss particular aspects of such representations: I will show how neural representations of segmentation masks allow enforcing constraints that are difficult to enforce in a classical pixel-wise segmentation, and illustrate how suitable neural representations of a video allow decomposing scenes in such a way that individual objects can be edited easily. Finally, I will discuss that the perspective of viewing images as functions leads to common image-to-image networks becoming neural operators, and illustrate different architectures for such neural operators at the example of improving the reconstruction of linear inverse problems.
Prof. Michael Möller
Bio: Michael Möller studied and completed his doctorate in applied mathematics in applied mathematics in Münster from 2004 to 2012, during which time he spent two years as a researcher at the University of California, Los Angeles, UCLA during this time. After completing his doctorate with his supervisor Martin Burger, he worked for 1.5 years in the research and development department department of Arnold & Richter Cine Technik GmbH (ARRI) before joining the computer vision group of Daniel Cremers at the Technical Cremers’ computer vision group at the Technical University of Munich. Since 2016 he is a professor at the University of Siegen. His main field of research is the combination of model- and learning-based methods for image reconstruction and analysis. He is part of the DFG priority program “Theoretical Foundations of Deep Learning”, spokesperson of the DFG research group “Learning to Sense” and was awarded a Lamarr Fellow in 2023.
Off the Grid - Neural Representations and Neural Operators for Image Processing
- Lamarr
Abstract: While images are classically represented on a regular grid of pixels, recent research has become increasingly interested in neural representations, i.e., the representation of scenes via parameterized continuous functions similar to neural networks. In this talk I will discuss particular aspects of such representations: I will show how neural representations of segmentation masks allow enforcing constraints that are difficult to enforce in a classical pixel-wise segmentation, and illustrate how suitable neural representations of a video allow decomposing scenes in such a way that individual objects can be edited easily. Finally, I will discuss that the perspective of viewing images as functions leads to common image-to-image networks becoming neural operators, and illustrate different architectures for such neural operators at the example of improving the reconstruction of linear inverse problems.
Prof. Michael Möller
Bio: Michael Möller studied and completed his doctorate in applied mathematics in applied mathematics in Münster from 2004 to 2012, during which time he spent two years as a researcher at the University of California, Los Angeles, UCLA during this time. After completing his doctorate with his supervisor Martin Burger, he worked for 1.5 years in the research and development department department of Arnold & Richter Cine Technik GmbH (ARRI) before joining the computer vision group of Daniel Cremers at the Technical Cremers’ computer vision group at the Technical University of Munich. Since 2016 he is a professor at the University of Siegen. His main field of research is the combination of model- and learning-based methods for image reconstruction and analysis. He is part of the DFG priority program “Theoretical Foundations of Deep Learning”, spokesperson of the DFG research group “Learning to Sense” and was awarded a Lamarr Fellow in 2023.