Tile2Vec: Unsupervised Representation Learning for Spatially Distributed Data

Neal Jean, Sherrie Wang, Anshul Samar, George Azzari, David B. Lobell, Stefano Ermon

AAAI 2019 pp. 3967-3974

doi:10.1609/AAAI.V33I01.33013967 /aaai/2019/jean2019aaai-tile/

Abstract

Geospatial analysis lacks methods like the word vector representations and pre-trained networks that significantly boost performance across a wide range of natural language and computer vision tasks. To fill this gap, we introduce Tile2Vec, an unsupervised representation learning algorithm that extends the distributional hypothesis from natural language — words appearing in similar contexts tend to have similar meanings — to spatially distributed data. We demonstrate empirically that Tile2Vec learns semantically meaningful representations for both image and non-image datasets. Our learned representations significantly improve performance in downstream classification tasks and, similarly to word vectors, allow visual analogies to be obtained via simple arithmetic in the latent space.

PDF AAAI Semantic Scholar

Cite

Text

Jean et al. "Tile2Vec: Unsupervised Representation Learning for Spatially Distributed Data." AAAI Conference on Artificial Intelligence, 2019. doi:10.1609/AAAI.V33I01.33013967

Markdown

[Jean et al. "Tile2Vec: Unsupervised Representation Learning for Spatially Distributed Data." AAAI Conference on Artificial Intelligence, 2019.](https://mlanthology.org/aaai/2019/jean2019aaai-tile/) doi:10.1609/AAAI.V33I01.33013967

BibTeX

@inproceedings{jean2019aaai-tile,
  title     = {{Tile2Vec: Unsupervised Representation Learning for Spatially Distributed Data}},
  author    = {Jean, Neal and Wang, Sherrie and Samar, Anshul and Azzari, George and Lobell, David B. and Ermon, Stefano},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2019},
  pages     = {3967-3974},
  doi       = {10.1609/AAAI.V33I01.33013967},
  url       = {https://mlanthology.org/aaai/2019/jean2019aaai-tile/}
}