Pith. sign in

REVIEW 7 cited by

Multi-Scale Representation Learning for Spatial Feature Distributions using Grid Cells

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2003.00824 v1 pith:PW7MZHKI submitted 2020-02-16 cs.CV cs.AIcs.LGstat.ML

Gengchen Mai , Krzysztof Janowicz , Bo Yan , Rui Zhu , Ling Cai , Ni Lao This is my paper
classification cs.CVcs.AIcs.LGstat.ML
keywords representationmulti-scalepositionsspatialtasksdifferentdistributionsmodel
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Unsupervised text encoding models have recently fueled substantial progress in NLP. The key idea is to use neural networks to convert words in texts to vector space representations based on word positions in a sentence and their contexts, which are suitable for end-to-end training of downstream tasks. We see a strikingly similar situation in spatial analysis, which focuses on incorporating both absolute positions and spatial contexts of geographic objects such as POIs into models. A general-purpose representation model for space is valuable for a multitude of tasks. However, no such general model exists to date beyond simply applying discretization or feed-forward nets to coordinates, and little effort has been put into jointly modeling distributions with vastly different characteristics, which commonly emerges from GIS data. Meanwhile, Nobel Prize-winning Neuroscience research shows that grid cells in mammals provide a multi-scale periodic representation that functions as a metric for location encoding and is critical for recognizing places and for path-integration. Therefore, we propose a representation learning model called Space2Vec to encode the absolute positions and spatial relationships of places. We conduct experiments on two real-world geographic data for two different tasks: 1) predicting types of POIs given their positions and context, 2) image classification leveraging their geo-locations. Results show that because of its multi-scale representations, Space2Vec outperforms well-established ML approaches such as RBF kernels, multi-layer feed-forward nets, and tile embedding approaches for location modeling and image classification tasks. Detailed analysis shows that all baselines can at most well handle distribution at one scale but show poor performances in other scales. In contrast, Space2Vec's multi-scale representation can handle distributions at different scales.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. POIFormer: A Transformer-Based Framework for Accurate and Scalable Point-of-Interest Attribution

    cs.LG 2025-07 conditional novelty 6.0 of 10

    A Transformer-based model that factorizes POI attribution into individual trajectory context and crowd-level spatiotemporal patterns, reporting accuracy gains over baselines on synthetic and real mobility datasets.

  2. Fine-Scale Soil Mapping in Alaska with Multimodal Machine Learning

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A multimodal deep learning system produces 10 m Alaska soil and permafrost maps and finds permafrost more sensitively than random forest under spatial holdout.

  3. Enhancing the KidSat Model: Integrating Geographical Encoding and Data Quality Assessment for Childhood Poverty Prediction

    cs.CV 2026-07 conditional novelty 5.0 of 10

    Refined DHS targets, two-stage image-quality screening, and spherical-harmonic geo-encoding reduce KidSat MAE from 0.2167 to 0.1759 (18.83 percent relative) and reach 0.1658 on 33 African countries.

  4. Omni Geometry Representation Learning vs Large Language Models for Geospatial Entity Resolution

    cs.DB 2025-08 unverdicted novelty 5.0 of 10

    A geometry-aware neural encoder plus attribute-aware language modeling improves geospatial entity resolution by up to 12% F1 over point-only baselines, with large language models competitive.

  5. Scalable Geospatial Data Generation Using AlphaEarth Foundations Model

    cs.LG 2025-08 conditional novelty 4.0 of 10

    A pipeline using AlphaEarth Foundations embeddings transfers US vegetation labels to Canada, reaching 73% accuracy on 13 vegetation classes.

  6. BERT4Traj: Transformer Based Trajectory Reconstruction for Sparse Mobility Data

    cs.LG 2025-07 conditional novelty 4.0 of 10

    A BERT-style transformer with demographic, anchor, and temporal context reconstructs missing locations in sparse mobility data more accurately than classical baselines on two Kampala datasets.

  7. Multi-Point Proximity Encoding For Vector-Mode Geospatial Machine Learning

    cs.LG 2025-06 conditional novelty 3.0 of 10

    MPP encoding, a distance-to-reference-points vectorization of arbitrary geospatial shapes, outperforms a raster indicator baseline at predicting shape properties and pairwise spatial relationships.

Pith tools