AI Paper Atlas

Research map / Probabilistic and scientific modeling / Molecules and proteins

TapSampling: Inference-Time Sampling with a Task-Progress-Understanding Verifier for Robotic Manipulation

Sizhe Zhao, Shengping Zhang, Shuo Yang, Weiyu Zhao, Shuigen Wang, Xiangyang Ji

ICML 2026 · Probabilistic and scientific modeling › Molecules and proteins · approach: Neuro-symbolic probabilistic layers · 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 TapSampling verifier-based samplingDiffusion Policy · CALVIN ABC->Daverage success length: 2.41 -> 2.58Improved
add TapSamplingOpenVLA · CALVIN ABC->Daverage success length: 3.30 -> 3.51Improved
add TapSamplingVPP · CALVIN ABC->Daverage success length: 4.39 -> 4.46Improved
add TapSamplingpi0.5 · LIBERO-Longaverage success rate: 96.8 -> 98.0Improved
add TapSamplingpi0 · real-world tasksaverage success rate: 78.3 -> 83.3Improved
use Learned Posterior Sampling instead of Gaussian Sampling (ablation)TapSampling on VPP · CALVIN ABC->Daverage success length: 4.43 -> 4.46Improved
use Learned Posterior Sampling instead of Gaussian Sampling (ablation)TapSampling on VPP · CALVIN ABC->D5th task success rate: 79.9 -> 81.1Improved
use Policy Sampling instead of Learned Posterior Sampling (ablation)TapSampling on VPP · CALVIN ABC->Daverage success length: 4.46 -> 4.50Improved
use Policy Sampling instead of Learned Posterior Sampling (ablation)TapSampling on VPP · CALVIN ABC->Dlatency (s): 0.488 -> 2.638Hurt
use Learned Posterior Sampling instead of Gaussian Sampling (ablation)TapSampling · action sampling MMD gamma=2.0MMD: 0.098 -> 0.064Improved
allow retry attempts (oracle) on VPP (ablation)VPP · CALVIN ABC->Daverage success length: 4.39 -> 4.72Improved
add TapSamplingpi0.5 · LIBERO-Longtask 1 and 8 success rate: 98.0 -> 96.0Mixed

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