AI Paper Atlas

Research map / Architectures and efficiency / Spiking networks

Efficient Transformer Attention for SNNs via Hadamard Simplification

Tingting Jiang, Jiangrong Shen, Long Chen, Yaxin Li, Qi Xu

ICML 2026 · Architectures and efficiency › Spiking networks · approach: Spiking transformers

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 patch splitting with 3x3 convolution (ablation)MM-Attn · CIFAR-100accuracy: 78.15% to 77.37%Hurt
replace MM-Attn with full Hadamard attention (ablation)MM-Attn (3x3 conv) · CIFAR-100accuracy: 77.37 -> 75.91Hurt
remove scale factor (ablation)FH-Attn · CIFAR-100accuracy: 75.91 -> 76.63Improved
remove multi-head attention (ablation)FH-Attn · DVS-Gestureaccuracy: 94.09 -> 95.48Improved
remove multi-head attention (ablation)FH-Attn · CIFAR-100accuracy: 76.63 -> 76.21No clear effect
use single mask on KV branch instead of double mask (ablation)SSA late-spiking · CIFAR-100accuracy: 76.21 -> 78.33Improved
use early-spiking instead of late-spiking gating (ablation)SSA single KV mask · CIFAR-100accuracy: 78.33 -> 79.45Improved
use early-spiking instead of late-spiking gating (ablation)SSA single KV mask · DVS-Gestureaccuracy: 96.52 -> 97.56Improved
compress K branch into channel-wise gate (USSA, 3x3 conv) (ablation)SSA · CIFAR-100accuracy: 79.45 -> 78.30Hurt
use 1x1 instead of 3x3 conv for K branch (ablation)USSA channel-wise · CIFAR-100accuracy: 78.30 -> 78.22Hurt
use channel-wise instead of spatial-wise reduction (ablation)USSA 1x1 conv · CIFAR-100accuracy: 77.84 -> 78.22Improved
replace MDSSA with SSASpikingResformer · CIFAR-100accuracy: 78.15 -> 79.45Improved

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