Physics Informed Machine Learning in Astronomy
Abstract:
Machine Learning had a significant success in Astronomy in recent years, and it becomes obvious that useful applications of ML require a tight connection to physical modelling. In this session, we will explore several aspects of imbuing physics as part of a ML model, from building hybrid models that merge both deep learning and physical models, to using known physical symmetries and equivariances to design dedicated neural architectures.
Organisers: Francois Lanusse (Chair) & Jean-Luc Starck (Committee member)
Session Schedule
- 17:30 - 17:50 Invited talk -
- Combining Physics and Machine Learning in Astronomy
- François Lanusse
- 17:50 - 18:10 Invited talk -
- Data-driven analysis of strong gravitational lenses: opportunities and challenges for machine learning
- Laurence Perreault-Levasseur
- 18:10 - 18:30 Invited talk -
- Multiscale modeling of galaxy-scale strong lenses
- Aymeric Galan*
- 18:30 - 18:50 Invited talk -
- Rethinking data-driven point spread function modeling with a differentiable optical model
- Tobias Liaudat*
- 18:50 - 19:10 Invited talk -
- Learning the Galaxy-Halo Connection by Imposing Exact Physical Symmetries
- Kate Storey-Fisher*
- 19:10 - 20:30 Invited poster -
- Sampling high-dimensional inverse problem posteriors with neural score estimation
- Benjamin Remy*
- 19:10 - 20:30 Contributed poster -
- Deep Learning-based galaxy image deconvolution
- Utsav Akhaury*
- 19:10 - 20:30 Contributed poster -
- Scalable and equivariant spherical CNNs by discrete-continuous (DISCO) convolutions
- Jason McEwen
- 19:10 - 20:30 Contributed poster -
- Delensing Gravitational Lensing Effects with Physics-Informed Neural Networks
- Ayoub Tajja*