PASC26

Presentation

From Visualization to Knowledge Discovery
DescriptionMachine learning–assisted analysis is now widespread in Astrophysics, but its scope is still narrow, mostly focusing on regression, classification, and anomaly detection tasks. On the other hand, the development of general visualization, exploration, and discovery tools has lagged behind. This session will present newly developed tools that enable scientific breakthroughs in Astrophysics and Cosmology by discovering unbiased, interpretable representations of data directly from the data and regardless of its type (e.g. images, spectra, datacubes, point clouds), and origin (i.e. from observations or sophisticated simulations). These tools automatically learn low-dimensional representations of complex objects such as simulated galaxies in multimodal data, and provide interactive visualization for exploring arbitrarily large datasets. The compact representations can be used downstream for model selection, simulation-based inference, hypothesis development, and experiment design. The tools are modular, easy to use, and require minimal coding experience, broadening access to complex datasets for both experts and the general public.
SlidesPDF
TimeTuesday, June 3014:45 – 15:15 CEST
LocationBldg. 6 – 102
Event Type
Minisymposium

Organised and co-sponsored by

Follow Us and Share