AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Diffuser: Efficient Transformers with Multi-Hop Attention Diffusion for Long Sequences

Aosong Feng, Irene Li, Yuang Jiang, Rex Ying

AAAI 2023 · Architectures and efficiency › Attention and transformers · approach: Sparse efficient transformers for long context · 7 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
Use multi-hop attention diffusion (ablation)Diffuser · LRAaverage accuracy: 52.36 -> 57.31Improved
Add local window attention (ablation)Diffuser · LRAaverage accuracy: 52.65 -> 57.31Improved
Add random attention (ablation)Diffuser · LRAaverage accuracy: 55.74 -> 57.31Improved
Add global attention (ablation)Diffuser · LRAaverage accuracy: 55.19 -> 57.31Improved
Replace randomly selected random attention with uniform distribution (ablation)Diffuser · LRAaverage accuracy: 57.31 -> 57.07No clear effect
Increase diffusion steps (ablation)Diffuser · Amazon-2048F1Improved
Use diffusion steps K≥5 (ablation)Diffuser · Amazon-2048F1No clear effect
Vary the teleport parameter α (ablation)Diffuser · Amazon-2048F1Mixed
Increase the number of local, global, or random attentions (ablation)Diffuser · LRAaccuracyImproved
Combine random and global attention at an optimal ratio (ablation)Diffuser · LRAaccuracyImproved

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