Research map / Vision-language and multimodal / Cross-modal learning
SMARTIES: Spectrum-Aware Multi-Sensor Auto-Encoder for Remote Sensing Images
Gencer Sumbul, Chang Xu, Emanuele Dalsasso, Devis Tuia
This paper tried a common change: Add mixup or CutMix
add cross-sensor token mixup · EuroSAT kNN · accuracy (%): 91.0 -> 93.2 Improved
What 547 accepted papers report for this change
- 416 improved
- 28 no clear effect
- 14 mixed
- 89 hurt
- Lightweight Federated Incremental Learning via Decoupled Replay ICML 2026 · add mixup to uploaded features · CIFAR100 (α=1.0) · A(f): 33.20 -> 35.61Improved
- Quantifying and Optimizing Simplicity via Polynomial Representations ICML 2026 · add embedding mixup · GLUE RTE, MRPC, CoLA · accuracy / Matthews corr.: MRPC 86.74 -> 86.30; CoLA 62.31 -> 61.04; RTE unchanged 70.28Hurt
- Fix the Mind, Not the Move: Interpretable AI Assistance via Knowledge-Gap Localization ICML 2026 · add auxiliary mixup generalization loss · Breakfast · System recall: 0.221 -> 0.762Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add cross-sensor token mixup (ablation) | SMARTIES ViT-B (50 epochs) · EuroSAT kNN | accuracy (%): 91.0 -> 93.2 | Improved |
| add cross-sensor token mixup with BEN-only pretraining (ablation) | SMARTIES ViT-B (50 epochs) · EuroSAT kNN | accuracy (%): 91.0 -> 91.1 | No clear effect |
| add cross-sensor token mixup (ablation) | SMARTIES · BEN-MM linear probe | mAP: +2.2% | Improved |
| use mixup concatenation instead of image stacking (ablation) | SMARTIES ViT-L · BEN-MM 10% linear probe | mAP: 83.1 -> 86.7 | Improved |
| use mixup concatenation instead of feature concatenation (ablation) | SMARTIES ViT-L · BEN-MM 1% linear probe | mAP: 77.0 -> 79.2 | Improved |
| add projection interpolation for unseen thermal bands (ablation) | SMARTIES ViT-B frozen · SICKLE | mIoU: 35.4 -> 50.2 | Improved |
| scale backbone from ViT-B to ViT-L (ablation) | SMARTIES · EuroSAT kNN | accuracy (%): 94.3 -> 94.6 | Improved |
| increase pretraining from 50 to 100 epochs (ablation) | SMARTIES ViT-B · EuroSAT kNN | accuracy (%): 93.2 -> 94.3 | Improved |
| add fMoW data to BEN pretraining (ablation) | SMARTIES ViT-B (50 epochs) · EuroSAT kNN | accuracy (%): 91.1 -> 93.2 | Improved |
| include SAR in pretraining versus projection extrapolation (ablation) | SMARTIES ViT-B (50 epochs) · BEN-S1 linear probe | mAP: 64.0 -> 73.6 | Improved |
| add multimodal input over SAR only (ablation) | SMARTIES ViT-L · BEN linear probe | mAP: 80.5 -> 86.7 | Improved |
| SMARTIES versus SatMAE++ on RESISC-45 | SatMAE++ (RGB) · RESISC-45 | top-1 accuracy (%): 97.5 -> 95.8 | Hurt |
Closest papers by meaning
- Galileo: Learning Global & Local Features of Many Remote Sensing Modalities (ICML 2025 · similarity 0.86)
- RoMA: Scaling up Mamba-based Foundation Models for Remote Sensing (NeurIPS 2025 · similarity 0.85)
- S2MAE: A Spatial-Spectral Pretraining Foundation Model for Spectral Remote Sensing Data (CVPR 2024 · similarity 0.84)
- A Scalable Multi-Task Model for Virtual Sensors (NeurIPS 2026 · similarity 0.84)
- Bridging Remote Sensors with Multisensor Geospatial Foundation Models (CVPR 2024 · similarity 0.84)
- Spiking Meets Attention: Efficient Remote Sensing Image Super-Resolution with Attention Spiking Neural Networks (NeurIPS 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
