GRSS Special Stream on Uncertainty-Aware and Robust Machine Learning for Remote Sensing
There has been a growing interest in the remote sensing community to develop machine/deep learning methods/models that are reliable and robust. Most existing models perform sub-optimally when they receive test samples drawn from outside the training distribution. For example, a model a trained on European urban areas may produce high confidence but erroneous predictions when tested on images from the south Asian urban areas. Anomalies and distribution shift are ubiquitous in the Earth observation data. To effectively deploy models under such scenario, it is crucial to either filter out out-of-distribution (OOD) samples by designing well-calibrated predictive uncertainty estimates or by generalizing well to the OOD samples by designing robust and domain generalizable models.
This IEEE GRSL Special Stream calls for uncertainty and robustness related contributions to the remote sensing domain. Possible topics include (but are not limited to):
- Out-of-distribution detection.
- Domain generalization.
- Domain adaptation.
- Confidence calibration and prediction.
- Spatial uncertainty.
- Temporal uncertainty.
- New benchmark datasets.
- Interpretability and explainability of machine learning models.
All submissions will be peer-reviewed according to the GRSS guidelines. Submitted articles should not have been published or be under review elsewhere. Submit your manuscript to mc.manuscriptcentral.com/grsl, using the Manuscript Central interface, and select our specific stream. Accepted papers are subject to GRSL’s usual publication charges.
- Dr. Sudipan Saha, Assistant Professor, Indian Institute of Technology, Delhi, India (Lead Guest Editor)
- Dr. Muhammad Shahzad, Guest Professor, Technical University of Munich, Germany (Guest Editor)
- Dr. Omid Ghorbanzadeh, Senior Researcher, IARAI, Austria (Guest Editor)
Submission open: July 1, 2023
Submission close: December 31, 2023
Paper Submission Link: mc.manuscriptcentral.com/grsl