AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

SaPaVe: Towards Active Perception and Manipulation in Vision-Language Action Models for Robotics

Mengzhen Liu, Enshen Zhou, Cheng Chi, Yi Han, Shanyu Rong, Liming Chen, Pengwei Wang, Zhongyuan Wang, Shanghang Zhang

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Memory and subgoal-based manipulation · 11 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
train specialized semantic camera policy (SaPaVe Stage 1) instead of using generalist VLMsemantic active perception · ActiveViewPose-200Ksuccess rate avg: 72.7 -> 84.3 (vs Gemini-2.5-Pro)Improved
use active head camera instead of fixed camera (ablation)SaPaVe architecture · ActiveManip-Benchavg success rate: 36.17 -> 74.83Improved
add wrist cameras to fixed head camera (ablation)fixed-camera model · ActiveManip-Benchavg success rate: 36.17 -> 52.33Improved
add wrist cameras to active head camera (ablation)SaPaVe · ActiveManip-Benchavg success rate: 74.83 -> 73.16No clear effect
use SaPaVe instead of fine-tuned GR00T-N1 / pi0real-world active manipulation · real-world 4 tasksavg success rate: 53.75 / 45.00 -> 85.00Improved
add Stage 1 semantic camera pretraining (ablation)SaPaVe · real-world 4 tasksavg success rate: 53.75 -> 85.00Improved
add Stage 2 active manipulation finetuning (ablation)SaPaVe · real-world 4 tasksavg success rate: 66.25 -> 85.00Improved
use decoupled action heads instead of a unified decoder (ablation)SaPaVe · real-world 4 tasksavg success rate: 71.25 -> 85.00Improved
use camera adapter (LoRA) instead of full finetuning (ablation)SaPaVe · real-world 4 tasksavg success rate: 73.75 -> 85.00Improved
add Universal Spatial Knowledge Injection (ablation)SaPaVe · real-world 4 tasksavg success rate: 68.75 -> 85.00Improved
test on unseen objects, lighting and scenesSaPaVe · real-world generalizationsuccess rate: 90 -> 85-95 (Occluded P&P)No clear effect

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