Research map / Architectures and efficiency / Attention and transformers
Learning Operators with Coupled Attention
Georgios Kissas, Jacob Seidman, Leonardo Ferreira Guilhoto, Victor M. Preciado, George J. Pappas, Paris Perdikaris
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add kernel-coupled attention (KCA) weights (ablation) | LOCA · Darcy flow, 1.5% labeled, N=200 | mean relative L2 error: 0.463 -> 0.017 | Improved |
| add KCA weights (ablation) | LOCA · Darcy flow, 1.5% labeled, N=200 | max relative L2 error: 3.023 -> 0.041 | Improved |
| LOCA vs DON and FNO with 6% labeled outputs | LOCA · Darcy flow | relative L2 error | Improved |
| LOCA vs DON and FNO at 2.5% labeled data | LOCA · Shallow Water | relative L2 error | Mixed |
| clean train, noisy test: LOCA vs FNO | LOCA · Mechanical MNIST | percent error increase v1: 1,930% -> 80% | Improved |
| clean train, noisy test: LOCA vs DeepONet | LOCA · Mechanical MNIST | percent error increase v1: 112% -> 80% | Improved |
| noisy train and test: LOCA vs FNO | LOCA · Mechanical MNIST | percent error increase v1: 280% -> 26% | Improved |
| noisy train and test: LOCA vs DeepONet | LOCA · Mechanical MNIST | percent error increase v1: 128% -> 26% | Improved |
| clean train and clean test: LOCA vs FNO | LOCA · Mechanical MNIST | mean relative L2 error v1: 0.004 -> 0.026 | Hurt |
| clean train and clean test: LOCA vs DeepONet | LOCA · Mechanical MNIST | mean relative L2 error v1: 0.028 -> 0.026 | No clear effect |
| different random initializations (10 seeds) | LOCA · Mechanical MNIST | max relative L2 error | Mixed |
| LOCA vs DON and FNO in temporal extrapolation | LOCA · Climate modeling, 4% labeled | relative L2 error | Improved |
Closest papers by meaning
- Nonlocal Attention Operator: Materializing Hidden Knowledge Towards Interpretable Physics Discovery (NeurIPS 2024 · similarity 0.80)
- Attention-based PCA (NeurIPS 2026 · similarity 0.79)
- MultiMax: Sparse and Multi-Modal Attention Learning (ICML 2024 · similarity 0.79)
- Improved Operator Learning by Orthogonal Attention (ICML 2024 · similarity 0.79)
- Learning Endogenous Attention for Incremental Object Detection (CVPR 2025 · similarity 0.79)
- In-Context Learning with Transformers: Softmax Attention Adapts to Function Lipschitzness (NeurIPS 2024 · similarity 0.79)
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.
