# Spatial Language Representation with Multi-Level Geocoding

arXiv: https://arxiv.org/abs/2008.09236 · HTML: https://mlsanity.com/paper/2008.09236/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 21 Aug 2020.

## Built on by (most cited first, 6 of 6)

- [A Review of Location Encoding for GeoAI: Methods and Applications](https://mlsanity.com/paper/2111.04006/) (2021)
- [There is a Time and Place for Reasoning Beyond the Image](https://mlsanity.com/paper/2203.00758/) (2022)
- [A Transformer-based Framework for POI-level Social Post Geolocation](https://mlsanity.com/paper/2211.01336/) (2022)
- [Regressing Location on Text for Probabilistic Geocoding](https://mlsanity.com/paper/2107.00080/) (2021)
- [Multilingual News Location Detection using an Entity-Based Siamese Network with Semi-Supervised Contrastive Learning and Knowledge Base](https://mlsanity.com/paper/2212.11856/) (2022)
- [Transformer Based Geocoding](https://mlsanity.com/paper/2301.01170/) (2023)

## Abstract (arXiv)

We present a multi-level geocoding model (MLG) that learns to associate texts to geographic locations. The Earth's surface is represented using space-filling curves that decompose the sphere into a hierarchy of similarly sized, non-overlapping cells. MLG balances generalization and accuracy by combining losses across multiple levels and predicting cells at each level simultaneously. Without using any dataset-specific tuning, we show that MLG obtains state-of-the-art results for toponym resolution on three English datasets. Furthermore, it obtains large gains without any knowledge base metadata, demonstrating that it can effectively learn the connection between text spans and coordinates - and thus can be extended to toponymns not present in knowledge bases.