AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

SOFT: Softmax-free Transformer with Linear Complexity

Jiachen Lu, Jinghan Yao, Junge Zhang, Xiatian Zhu, Hang Xu, Weiguo Gao, Chunjing Xu, Tao Xiang, Li Zhang

NeurIPS 2021 · Architectures and efficiency › Attention and transformers · approach: Attention head and block redesign · 59 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
Replace softmax dot-product attention with Gaussian-kernel SOFT attentionTiny Transformer · ImageNet-1K validationTop-1 accuracy (%): 78.2 -> 79.3Improved
Replace the PVT attention block with SOFT attentionPVT-style models · ImageNet-1K validation, model variantsTop-1 accuracy (%): 75.1 -> 79.3; 79.8 -> 82.9; 81.2 -> 82.9; 81.7 -> 83.1Improved
Use SOFT on LRA tasks instead of NyströmformerEfficient Transformer · LRA Retrieval (4K)score: 79.56 -> 81.77Improved
Use SOFT instead of NyströmformerEfficient Transformer · LRA Image (1K)score: 41.58 -> 46.91Improved
Use SOFT instead of NyströmformerEfficient Transformer · LRA Text (4K)score: 65.52 -> 63.49Hurt
Use bottleneck length 64 instead of 49 (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 79.3 -> 79.3No clear effect
Use bottleneck length 81 instead of 49 (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 79.3 -> 78.9Hurt
Use bottleneck length 36 instead of 49 (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 79.3 -> 79.0Hurt
Use biased sampling instead of convolution sampling (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 79.3 -> 79.0Hurt
Use average pooling instead of convolution sampling (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 79.3 -> 79.3No clear effect
Add a pyramidal architecture (ablation)SOFT-Small · ImageNet-1K validationTop-1 accuracy (%): 80.1 -> 82.2Improved
Add overlapping convolution (ablation)SOFT-Tiny · ImageNet-1K validationTop-1 accuracy (%): 77.4 -> 79.3Improved

Later papers that compared with this method

2 later accepted papers report SOFT++-Medium as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.

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