Self-supervised Pretraining of Vision Transformers for Earth Observation
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Remote sensing offers vast yet sparsely labeled multimodal data but lacks foundation models that can be leveraged across societally impactful applications. In this thesis, I develop foundation models for Earth Observation by pretraining vision transformers (ViTs) using self-supervised learning. Using masked autoencoding, I develop one of the first ViTs pretrained on RS data—called SatViT. Next, I develop an improved foundation model—called SatViT-V2—and evaluate it on out-of-distribution data. I show that these models are more robust to distribution shifts than models typically used by the remote sensing community. Next, I develop a novel framework that learns SoTA representations for Earth Observation—called CROMA—by jointly leveraging cross-modal contrastive learning and multimodal masked autoencoding. In CROMA, I also extend a SoTA position encoding method to cross-attention and ViTs—called X- and 2D-ALiBi—and demonstrate their superior accuracy and flexibility. My thesis shows that suitably pretrained ViTs can form effective foundation models for Earth Observation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it