AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Adaptive Physics Transformer with Fused Global-Local Attention for Subsurface Energy Systems

Xin Ju, Hadrian Fung, Yuyan Zhang, Carl Jacquemyn, Matthew D Jackson, Randolph R. Settgast, Sally M Benson, Gege Wen

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Transformer neural operators for PDEs · 0 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
fuse global perceiver and local GNO branches with gating instead of global only (ablation)APT · 2D faulted CO2 storagesaturation error dSg (%): 1.50 -> 0.32Improved
add global branch to local-only GNO encoder via fusion (ablation)APT · 2D faulted CO2 storagepressure error dPg (%): 0.18 -> 0.11Improved
add gated fusion (ablation)APT · ATES adaptive meshrelative L2: 0.014 -> 0.011Improved
add fused attention to GNO-only variant (ablation)APT · Basin-scale CO2 storage with LGRsaturation error (%): 3.9% -> 2.5%Improved
add gated fusion (ablation)APT · CarBench super-resolutionscore at full resolution: 95.2 -> 96.5Improved
add gated fusion (ablation)APT · ATES super-resolutionR2 at full resolution (%): 96.5 -> 97.2Improved
query at higher resolution than training (ATES)APT (fused) · ATES super-resolutionR2 (%): 99.0 -> 97.2Hurt
query at higher resolution than training (CarBench)APT (fused) · CarBench super-resolutionscore: 96.0 -> 96.5Improved
joint training with Gaussian LGR dataset (ablation)APT · Channelized CO2 storageR2 (%): 32.0 -> 86.7Improved
increase training samples from 40k to 161k (ablation)APT · ATES data scalingtest loss: 0.1153 -> 0.0812Improved
increase model size from 0.8M to 69.4M parameters (ablation)APT · ATES model scalingtest loss: 0.1167 -> 0.0722Improved
APT versus U-FNO on hydrocarbon pressureAPT · Cartesian hydrocarbon (40x40)pressure error dP (%): 0.57 -> 0.60No clear effect

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