AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

A Hierarchical Spatial Transformer for Massive Point Samples in Continuous Space

Wenchong He, Zhe Jiang, Tingsong Xiao, Zelin Xu, Shigang Chen, Ronald Fick, MILES D MEDINA, Christine Angelini

NeurIPS 2023 · Architectures and efficiency › Attention and transformers · approach: Efficient image transformers · 3 citations (OpenAlex)

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 changedWhereMeasuredEffect
add neural feature learning to GP (Deep GP vs GP)GP · Red tideMSE: 7.42 -> 6.23Improved
use hierarchical quadtree attention (HST) instead of GNN message passing (Spatial GNN)Spatial GNN · Red tideMSE: 5.68 -> 5.25Improved
replace fixed-graph GNN with all-pair attentionSpatial GNN · Red tideMSE: 5.68 -> 5.30Improved
multi-level message passing (MGNO) over spatial GNNSpatial GNN · Red tideMSE: 5.68 -> 5.41Improved
HST vs all-pair transformerAll-pair transformer · Red tideMSE: 5.30 -> 5.25Improved
HST vs all-pair transformerAll-pair transformer · TurbidityMSE: 0.31 -> 0.30Improved
HST vs all-pair transformerAll-pair transformer · Darcy flowMSE: 0.10 -> 0.11Hurt
HST vs Galerkin TransformerGalerkinTransformer · Darcy flowMAE: 0.25 -> 0.23Mixed
HST uncertainty branch vs GPGP · Red tide UQAvU: 0.33 -> 0.54Improved
HST uncertainty branch vs Spatial GNNSpatial GNN · Red tide UQAvU: 0.40 -> 0.54Improved
HST uncertainty branch vs Galerkin TransformerGalerkinTransformer · Red tide UQAvU: 0.38 -> 0.54Improved
quadtree leaf size threshold 50 to 200HST · computation time simulationcomputation timeHurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy