The {Champagne} model: {Cosmic} {Void} {Detection} based on marked point processes.

@inproceedings{Noerby2026RM,
 abstract = {The large-scale structure of the Universe, known as the cosmic web, exhibits a wide variety of environments, including filaments, walls, clusters, and voids. Current and next-generation cosmological surveys are set to dramatically increase the resolution, depth, and number density of observed galaxies, enabling detailed statistical analyses of large populations of cosmic voids. Void detection and characterisation from galaxy distributions constitute a pattern recognition problem that can be tackled with marked point processes with interactions, as has previously been done for the filament and cluster components of the cosmic web. Here, we present a new marked Gibbs point process model with interactions designed to detect the void pattern in the cosmic web, that is, to trace underdense regions in the spatial galaxy distribution. The model assumes the void pattern is made of interacting overlapping bubbles. The model consists of two components: a data term that controls the placement of bubbles within underdense regions of the observed galaxy distribution, and an interaction term that governs the geometry of bubbles overlap and interaction. The two components acting together map the void pattern. Simulation of the model is performed via a Metropolis-Hastings algorithm. The model and the dynamical parameters are chosen using ad hoc strategies based on physical and numerical arguments. Samples obtained from the previously fitted model are used to construct a visit map that approximates the probability that each point in the observed domain belongs to the void pattern. Finally the sufficient statistics of the model and the visit map, along with cosmological arguments, are used to verify the obtained results. Results are presented for both simulated and real data, demonstrating the model’s ability to recover cosmic void structures through a stochastic approach to void detection.},
 author = {Noerby, Freja Amalie and Stoica, Radu and Gemmerlé, Didier and van de Weygaert And Tempel, Elmo},
 booktitle = {2026 {RING} meeting},
 language = {en},
 pages = {87--107},
 publisher = {ASGA},
 title = {The {Champagne} model: {Cosmic} {Void} {Detection} based on marked point processes.},
 year = {2026}
}