AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

EMFormer: Efficient Multi-Scale Transformer for Accumulative Context Weather Forecasting

Hao Chen, Tao Han, Jie ZHANG, Song Guo, Fenghua Ling, LEI BAI

ICML 2026 · Architectures and efficiency › Attention and transformers · approach: Transformer expressivity and circuits · 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
add accumulative context finetuning (ablation)EMFormer · ERA5 1.4 deg, 10-dayRMSE: 0.5719 -> 0.5094Improved
add accumulative context finetuning (ablation)EMFormer · ERA5 1.4 deg, 10-dayACC: 0.4614 -> 0.5389Improved
add KV pruning memory module to accumulative finetuning (ablation)EMFormer · ERA5 1.4 deg, 10-dayRMSE: 0.5153 -> 0.5094Improved
add KV pruning memory module (ablation)EMFormer · ERA5 1.4 degGPU memory (G): 4.576 -> 4.913Hurt
add plain autoregressive finetuning (no accumulation) (ablation)EMFormer · ERA5 1.4 deg, 6-hourRMSE: 0.0626 -> 0.0645Hurt
add plain autoregressive finetuning (no accumulation) (ablation)EMFormer · ERA5 1.4 deg, 10-dayRMSE: 0.5719 -> 0.5339Improved
use LoRA finetuning instead of full finetuning (ablation)EMFormer · ERA5 1.4 deg, 10-dayRMSE: 0.5339 -> 0.5278Mixed
use EMFormer block instead of self-attention block (ablation)EMFormer backbone · ERA5 1.4 deg, 10-dayRMSE: 0.5978 -> 0.5719Improved
use EMFormer block instead of self/windows attention block (ablation)EMFormer backbone · ERA5 1.4 deg, 10-dayRMSE: 0.6016 -> 0.5719Improved
use multi-convs layer instead of plain multi-scale module (ablation)EMFormer · ERA5 weather, 6-hourtraining time (hours): 83 -> 60Improved
use multi-convs layer instead of plain multi-scale module (ablation)EMFormer-B · ImageNet-1KTop-1 accuracy (%): 84.3 -> 84.4No clear effect
use multi-convs layer instead of plain multi-scale module (ablation)EMFormer · ERA5 weather, 6-hour Z500RMSE: 18.6 -> 18.6No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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