Research map / Architectures and efficiency / Attention and transformers
FMMformer: Efficient and Flexible Transformer via Decomposed Near-field and Far-field Attention
Tan Minh Nguyen, Vai Suliafu, Stanley Osher, Long Chen, Bao Wang
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 Band5 near-field attention to linear transformer (1 kernel) (ablation) | Linear transformer · LRA | average test accuracy: 54.67 -> 59.51 | Improved |
| add second feature map (2 kernels) to FMMformer (ablation) | FMMformer + Band5 · LRA | average test accuracy: 59.51 -> 60.74 | Improved |
| add low-rank far-field attention to banded attention (ablation) | Band5 · LRA | average test accuracy: 57.73 -> 59.51 | Improved |
| add near-field Band5 to linear transformer on ListOps (ablation) | Linear transformer · LRA ListOps | test accuracy: 18.30 -> 33.22 | Improved |
| add near-field Band5 to linear transformer on Pathfinder (ablation) | Linear transformer · LRA Pathfinder | test accuracy: 71.17 -> 71.29 | No clear effect |
| add linear far-field to Band5 on Retrieval (ablation) | Band5 · LRA Retrieval | test accuracy: 79.41 -> 81.50 | Improved |
| add Band5 near-field to linear transformer (ablation) | Linear transformer · WikiText-103 | test perplexity: 38.40 -> 37.29 | Improved |
| increase bandwidth from 5 to 20 (ablation) | FMMformer 1-kernel · WikiText-103 | test perplexity: 37.29 -> 36.43 | Improved |
| use 2 kernels instead of 1 with Band20 (ablation) | FMMformer · WikiText-103 | test perplexity: 36.43 -> 36.11 | Improved |
| replace linear far-field with fast weight update (2 kernels) (ablation) | FMMformer + Band20 · WikiText-103 | test perplexity: 36.11 -> 34.71 | Improved |
| replace linear far-field with fast weight update (1 kernel) (ablation) | FMMformer + Band20 · WikiText-103 | test perplexity: 36.43 -> 35.47 | Improved |
| add far-field low-rank attention to banded Band20 (ablation) | Band20 · WikiText-103 | test perplexity: 39.19 -> 36.43 | Improved |
Closest papers by meaning
- PDE-Transformer: Efficient and Versatile Transformers for Physics Simulations (ICML 2025 · similarity 0.88)
- SEA: State-Exchange Attention for High-Fidelity Physics Based Transformers (NeurIPS 2024 · similarity 0.86)
- Universal Approximation of Mean-Field Models via Transformers (ICML 2025 · similarity 0.85)
- Erwin: A Tree-based Hierarchical Transformer for Large-scale Physical Systems (ICML 2025 · similarity 0.84)
- A multiscale analysis of mean-field transformers in the moderate interaction regime (NeurIPS 2025 · similarity 0.84)
- Transolver: A Fast Transformer Solver for PDEs on General Geometries (ICML 2024 · similarity 0.84)
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.
