Research map / Label-efficient and robust learning / Recognition and long-tail
Tile2Vec: Unsupervised Representation Learning for Spatially Distributed Data
Neal Jean, Sherrie Wang, Anshul Samar, George Azzari, David Lobell, Stefano Ermon
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| learn embeddings from geographically neighboring and distant tile triplets | Tile2Vec on CDL land cover classification · CDL land cover classification, n=1000 | accuracy: Tile2Vec: 52.6 ± 1.1; Autoencoder: 49.1 ± 0.7 | Improved |
| learn embeddings from geographically neighboring and distant tile triplets | Tile2Vec on CDL land cover classification · CDL land cover classification, n=10000 | accuracy: Tile2Vec: 59.7 ± 0.3; Autoencoder: 55.6 ± 0.2 | Improved |
| learn Tile2Vec embeddings from Landsat imagery | Uganda poverty prediction · Uganda LSMS | r^2: 0.41 -> 0.496 ± 0.014 | Improved |
| learn Tile2Vec embeddings from non-health country features using spatial triplets | Country health index prediction · CIA Factbook countries | r^2: kNN: 62.8 ± 1.5 -> 77.5 ± 1.0; RF: 72.1 ± 1.6 -> 76.0 ± 1.3; RR: 68.7 ± 1.7 -> 69.6 ± 1.0 | Improved |
| use neighborhood radius 100 instead of sampling tiles without a neighborhood constraint (ablation) | Tile2Vec, tile size 100, LR · CDL land cover classification | accuracy: 35.2 ± 0.0 -> 61.8 ± 0.4 | Improved |
| increase neighborhood radius from 100 to 1000 pixels (ablation) | Tile2Vec on NAIP · CDL land cover classification | accuracy: 61.8 ± 0.4 -> 55.1 ± 0.4 | Hurt |
| vary the triplet-loss margin from 0.1 to 100 (ablation) | Tile2Vec on NAIP · CDL land cover classification | accuracy | No clear effect |
| remove embedding-magnitude regularization (ablation) | Tile2Vec · CDL land cover classification | feature quality | No clear effect |
| increase tile size (ablation) | Tile2Vec on NAIP · CDL land cover classification | accuracy: At neighborhood 100, LR: 57.8 ± 0.4 -> 61.8 ± 0.4 (tile size 25 to 100) | Improved |
Later papers that compared with this method
1 later accepted paper reports Tile2Vec as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Measuring Fine-Grained Urban Air Temperature with Satellite Imagery (AAAI 2025 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Textual Supervision Enhances Geospatial Representations in Vision-Language Models (ICML 2026 · similarity 0.88)
- GeoDiT: A Diffusion-based Vision-Language Model for Geospatial Understanding (CVPR 2026 · similarity 0.85)
- Poly2Vec: Polymorphic Fourier-Based Encoding of Geospatial Objects for GeoAI Applications (ICML 2025 · similarity 0.85)
- TorchSpatial: A Location Encoding Framework and Benchmark for Spatial Representation Learning (NeurIPS 2024 · similarity 0.85)
- Not Only Text: Exploring Compositionality of Visual Representations in Vision-Language Models (CVPR 2025 · similarity 0.84)
- Geolocation Representation from Large Language Models Are Generic Enhancers for Spatio-Temporal Learning (AAAI 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.
