| First Authors |
Dominik Sturm, Hiba Bensalem
|
| Authors |
Dominik Sturm, Hiba Bensalem, Ivo F. Sbalzarini
|
| Corresponding Authors |
Ivo F. Sbalzarini
|
| Last Authors |
Ivo F. Sbalzarini
|
| Conference Proceedings Volume Title |
International Conference on Learning Representations (ICLR) |
| Series Title |
|
| Conference Name |
International Conference on Learning Representations (ICLR) |
| Conference Location |
Rio de Janeiro, Brazil |
| Conference Start Date |
|
| Conference End Date |
|
| Chapter Number |
|
| Publisher |
ICLR |
| Conference Proceedings Editors |
|
| ISBN |
|
| First Page |
1 |
| Last Page |
35 |
| Open Access |
true |
| Print Publication Date |
2026-01-01 |
| Online Publication Date |
|
| Abstract |
We introduce spatially informed variational autoencoders (SI-VAE) as self-supervised deep-learning models that use stochastic point processes to predict spatial organization patterns from images. Existing approaches to learning visual representations based on variational autoencoders (VAE) struggle to capture spatial correlations between objects or events, focusing instead on pixel intensities. We address this limitation by incorporating a point-process likelihood, derived from the Papangelou conditional intensity, as a self-supervision target. This results in a hybrid model that learns statistically interpretable representations of spatial localization patterns and enables zero-shot conditional simulation directly from images. Experiments with synthetic images show that SI-VAE improve the classification accuracy of attractive, repulsive, and uncorrelated point patterns from 48% (VAE) to over 80% in the worst case and 90% in the best case, while generalizing to unseen data. We apply SI-VAE to a real-world microscopy data set, demonstrating its use for studying the spatial organization of proteins in human cells and for using the representations in downstream statistical analysis. |
| PDF |
Sturm_2026_9155.pdf
(28.4 MB)
|
| Cover Image |
|
| Affiliated With |
Sbalzarini
|
| Selected By |
Sbalzarini
|
| Acknowledged Services |
|
| Publication Status |
Published |
| Edoc Link |
|
| Sfx Link |
|
| DOI |
https://openreview.net/forum?id=09YSBymX6O |
| PubMed ID |
|
| WebOfScience Link |
|
| Alternative Full Text URL |
https://openreview.net/pdf?id=09YSBymX6O |
| Display Publisher Download Only |
false |
| Visible On MPI-CBG Website |
true |
| PDF Downloadable |
true |
| Created By |
sbalzari |
| Added Date |
2026-02-25 |
| Last Edited By |
sbalzari |
| Last Edited Date |
2026-02-25 14:46:27.582 |
| Library ID |
9155 |
| Document ID |
|
| Entry Complete |
false |
| eDoc Compliant |
false |
| Include in Edoc Report |
false |
| In Pure |
false |
| Ready for eDoc Export |
false |
| Author Affiliations Complete |
false |
| Project Name |
ScaDS.AI |
| Project URL |
|
| Grant ID |
|
| Funding Programme |
|
| Funding Organisation |
BMFTR |