Search papers, labs, and topics across Lattice.
The paper introduces Trust-SSL, a self-supervised learning framework designed to improve robustness to corruptions in aerial imagery by incorporating a per-sample, per-factor trust weight into the contrastive loss objective as an additive residual. This approach avoids enforcing alignment between clean and severely degraded views, which can introduce spurious structure. Experiments on EuroSAT, AID, and NWPU-RESISC45 datasets demonstrate that Trust-SSL achieves state-of-the-art linear probe accuracy and significant improvements under severe corruptions, particularly haze, while also showing consistent gains in zero-shot cross-domain stress tests.
Stop punishing your model for disagreeing with corrupted data – Trust-SSL learns better representations by treating alignment with degraded views as a residual learning problem, not a hard constraint.
Self-supervised learning (SSL) is a standard approach for representation learning in aerial imagery. Existing methods enforce invariance between augmented views, which works well when augmentations preserve semantic content. However, aerial images are frequently degraded by haze, motion blur, rain, and occlusion that remove critical evidence. Enforcing alignment between a clean and a severely degraded view can introduce spurious structure into the latent space. This study proposes a training strategy and architectural modification to enhance SSL robustness to such corruptions. It introduces a per-sample, per-factor trust weight into the alignment objective, combined with the base contrastive loss as an additive residual. A stop-gradient is applied to the trust weight instead of a multiplicative gate. While a multiplicative gate is a natural choice, experiments show it impairs the backbone, whereas our additive-residual approach improves it. Using a 200-epoch protocol on a 210,000-image corpus, the method achieves the highest mean linear-probe accuracy among six backbones on EuroSAT, AID, and NWPU-RESISC45 (90.20% compared to 88.46% for SimCLR and 89.82% for VICReg). It yields the largest improvements under severe information-erasing corruptions on EuroSAT (+19.9 points on haze at s=5 over SimCLR). The method also demonstrates consistent gains of +1 to +3 points in Mahalanobis AUROC on a zero-shot cross-domain stress test using BDD100K weather splits. Two ablations (scalar uncertainty and cosine gate) indicate the additive-residual formulation is the primary source of these improvements. An evidential variant using Dempster-Shafer fusion introduces interpretable signals of conflict and ignorance. These findings offer a concrete design principle for uncertainty-aware SSL. Code is publicly available at https://github.com/WadiiBoulila/trust-ssl.
