| Pushing the Boudaries of Structure-from-Motion with Machine Learning
Eric Brachmann, Scientist at Niantic Spatial
In 3D computer vision, we are currently witnessing a remarkable renaissance of interest in structure-from-motion (SfM), i.e. estimating camera poses and 3D geometry from a collection of images. Of course, SfM was never gone. Rather, solutions based on feature-matching and traditional multi-view geometry matured to a state about 10 years ago that turned them into reliable off-the-shelf components for various 3D vision tasks. Still, traditional SfM approaches are most reliable when certain conditions are met. For example, reconstructing very few or a huge amount of images can be challenging. The talk will investigate how learning-based formulations of SfM can address these challenges. We will focus on scene coordinate regression, an implicit scene representation, that naturally avoids the explosion of complexity inherent to image-to-image matching when the number of images is large. The talk culminates in the presentation of ACEZero, a self-supervised scene coordinate regression pipeline, that is able to reconstruct 10.000 images in reasonable time. | | | |
| European Conference on Numerical Mathematics and Advanced Applications (ENUMATH 2025)
The European Conference on Numerical Mathematics and Advanced Applications (ENUMATH) conferences are a forum for presenting and discussing novel and fundamental advances in numerical mathematics and challenging scientific and industrial applications on the highest level of international expertise. They started in Paris in 1995. Subsequent ENUMATH conferences were held at the universities of Heidelberg (1997), Jyväskylä (1999), Ischia Porto (2001), Prague (2003), Santiago de Compostela (2005), Graz (2007), Uppsala (2009), Leicester (2011), Lausanne (2013), Ankara (2015),
Bergen (2017), Egmond aan Zee (2019) and Lisabon (2023).
Conference Themes
- Advances in Discretisation Schemes
- Multiscale and Multiphysics Problems
- Hardware-Aware Scientific Computing
- Inverse Problems
- Uncertainty Quantification
- Data-Driven Modelling and Simulation
- Scientific Machine Learning
- Reduced Order Models and Surrogates
- Randomised Numerical Algorithms
- Numerical Optimisation and Optimal Control
| Local Organizing Committee
- Robert Scheichl (Chair) - IMa and IWR, Heidelberg University
- Peter Bastian - IWR, Heidelberg University
- Roland Herzog - IWR, Heidelberg University
- Vincent Heuveline - IWR, Heidelberg University
- Guido Kanschat - IWR, Heidelberg University
- Ekaterina Kostina - IWR, Heidelberg University
- Jakob Zech - IWR, Heidelberg University
| Neuenheimer Feld” campus of Heidelberg University in the Mathematikon building (Im Neuenheimer Feld 205), the Centre for Organismal Studies (Im Neuenheimer Feld 230) and the Chemistry lecture halls (Im Neuenheimer Feld 252). | |
| IWR School on Machine Learning for Fundamental Physics
Machine Learning is here to stay and is shaping the future of fundamental physics research. From optimal inference, over theory-inspired network architectures, to anomaly detection, representation learning and foundation models, a new generation of scientists is driving these exciting developments. This school aims to further strengthen technical expertise and foster new connections.
The 2025 IWR School on Machine Learning for Fundamental Physics is aimed at advanced PhD students specializing in scientific machine learning. We particularly encourage registrations from researchers with experience in scientific machine learning, as demonstrated by papers or preprints related to the topic of the school. The school takes place at the Interdisciplinary Center for Scientific Computing (IWR) at Heidelberg University September 15th-19th 2025. | Organizers:
Scientific: Tilman Plehn, David Shih, Caroline Heneka
At IWR: Jan Keese, Anne Leonhardt, Michael Winckler | IWR, Mathematikon, Im Neuenheimer Feld 205, 69120 Heidelberg | |
| International Marsilius Academy:
„AI and Human Values: Exploring technological, social, and normative perspectives“
About the event:
The rapid development of generative AI is having a profound impact on modern society. At the same time, it raises critical questions about the underlying values and the ethical, legal, and societal implications of these technologies. The aim of the Summer School is to critically examine the normative foundations of generative AI and to discuss which values are embedded in its design and how these values can be reflected upon and modeled. The Summer School offers an interdisciplinary space to connect perspectives from computer science, linguistics, philosophy, theology, medicine, and law. It is aimed at early-career researchers who seek to engage in meaningful discourse on values and norms in AI development and wish to build lasting scholarly networks.
Application deadline: June 27, 2025 | Organisation:
Marsilius-Kolleg und Heidelberg Center for Digital Humanities
Wissenschaftliches Komitee:
Maria Becker (Computerlinguistik),
Michael Boutros (Genomforschung),
Michael Gertz (Informatik),
Nora Heinzelmann (Philosophie),
Friederike Nüssel (Theologie) | Marsilius Kolleg, Heidelberg
INF130.1 | |
