Abstract
We propose a new method for generating realistic datasets with distribution shifts using any decoder-based generative model. Our approach systematically creates datasets with varying intensities of distribution shifts, facilitating a comprehensive analysis of model performance degradation. We then use these generated datasets to evaluate the performance of various commonly used networks and observe a consistent decline in performance with increasing shift intensity, even when the effect is almost perceptually unnoticeable to the human eye. We see this degradation even when using data augmentations. We also find that enlarging the training dataset beyond a certain point has no effect on the robustness and that stronger inductive biases increase robustness.
| Original language | English |
|---|---|
| Title of host publication | Computer Vision – ECCV 2024 Workshops, Proceedings |
| Editors | Alessio Del Bue, Cristian Canton, Jordi Pont-Tuset, Tatiana Tommasi |
| Publisher | Springer Science and Business Media Deutschland GmbH |
| Pages | 321-334 |
| Number of pages | 14 |
| ISBN (Print) | 9783031919060 |
| DOIs | |
| State | Published - 2025 |
| Event | 18th European Conference on Computer Vision, ECCV 2024 - Milan, Italy Duration: 29 Sep 2024 → 4 Oct 2024 |
Publication series
| Name | Lecture Notes in Computer Science |
|---|---|
| Volume | 15642 LNCS |
| ISSN (Print) | 0302-9743 |
| ISSN (Electronic) | 1611-3349 |
Conference
| Conference | 18th European Conference on Computer Vision, ECCV 2024 |
|---|---|
| Country/Territory | Italy |
| City | Milan |
| Period | 29/09/24 → 4/10/24 |
Bibliographical note
Publisher Copyright:© The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
Fingerprint
Dive into the research topics of 'Control+Shift: Generating Controllable Distribution Shifts'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver