AI Paper Atlas

Research map / Architectures and efficiency / Spiking networks

SmoothSpike: Spiking Transformer with Learnable Hadamard Transformation

Zijian Zhou, Wenjie Wei, Yu Liang, Jialin Li, Ammar Belatreche, Honglin Cao, Shuai Wang, Malu Zhang, Yang Yang, Haizhou Li

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
add SmoothSpike to SpikingformerSpikingformer · GLUE devaverage score: 66.8 -> 75.0Improved
add SmoothSpike to SpikeLMSpikeLM · GLUE devaverage score: 75.7 -> 77.5Improved
add SmoothSpike to SpikeLMSpikeLM · GLUE RTEscore: 65.3 -> 58.1Hurt
add SmoothSpike to SpikingformerSpikingformer · GLUE STS-Bscore: 54.5 -> 79.3Improved
add SmoothSpike to SpikingformerSpikingformer · GLUE CoLAscore: 24.4 -> 41.2Improved
add SmoothSpike to SpikingformerSpikingformer · GLUE energyenergy (mJ): 6.76 -> 9.45Hurt
add learnable Hadamard H1 to Spikingformer (ablation)Spikingformer baseline · GLUE dev ablationaverage score: 64.00 -> 65.88Improved
add learnable H2 after H1 (ablation)Spikingformer + learnable H1 · GLUE dev ablationaverage score: 65.88 -> 67.64Improved
add learnable H3 after H1 and H2 (ablation)Spikingformer + learnable H1,H2 · GLUE dev ablationaverage score: 67.64 -> 67.88Improved
add fixed randomized Hadamard H1 instead of nothing (ablation)Spikingformer baseline · GLUE dev ablationaverage score: 64.00 -> 63.74No clear effect
make H1 learnable instead of fixed (ablation)Spikingformer + H1 · GLUE dev ablationaverage score: 63.74 -> 65.88Improved
apply H3 learnable Hadamard transform (ablation)Spikingformer · saturated neuron rate (Layer 4)saturation rate (%): 3.79 -> 0.60Improved

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