AI Paper Atlas

Research map / Architectures and efficiency / Pruning and efficient tuning

SpecPrune-VLA: Accelerating Vision-Language-Action Models via Action-Aware Self-Speculative Pruning

Hanzhen Wang, Jiaming Xu, Yushun Xiang, Jiayi Pan, Yongkang Zhou, Yong-Lu Li, Guohao Dai

ICML 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Layer-wise pruning of LLMs · 0 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
add global attention reuse from the previous step to static pruning (ablation)SpecPrune-VLA · LIBEROsuccess rate (%): 93.4% -> 96.1%Improved
use entropy-based layer weighting instead of average weighting in dynamic pruning (ablation)SpecPrune-VLA · LIBEROsuccess rate (%): 92.0% -> 96.1%Improved
add the action-aware controller on top of static and dynamic pruning (ablation)SpecPrune-VLA · LIBERO-Spatialsuccess rate (%): 96.8 -> 97.4Improved
add action-aware controller (ablation)SpecPrune-VLA · LIBERO-Spatiallatency: 70.8ms -> 72.3msMixed
guide pruning with mid and deep layer attention instead of shallow layers (ablation)post-hoc token pruning on OpenVLA-OFT · LIBERO pick and placesuccess rateImproved
random token pruning beyond 12.5% sparsity (ablation)OpenVLA-OFT · LIBERO pick and placesuccess rateHurt
use first two layers rather than first layer for local speculation (ablation)SpecPrune-VLA · LIBEROhit rateImproved
add a third layer for local speculation (ablation)SpecPrune-VLA · LIBEROhit rate and latencyNo clear effect
lower prune ratio alpha (more tokens pruned) (ablation)SpecPrune-VLA · LIBEROsuccess rate vs speedupMixed
apply SpecPrune-VLA instead of FastV at similar token countsOpenVLA-OFT · LIBEROaverage success rate (%): 93.3% -> 96.1%Improved
apply SpecPrune-VLA to OpenVLA-OFTOpenVLA-OFT · LIBEROaverage speedup: 1.00× -> 1.46×Improved
apply SpecPrune-VLA on real robotOpenVLA-OFT · Real-world taskslatency (ms): 187.7 -> 110.2Improved

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