AI Paper Atlas

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

Spatially Guided Training for Vision-Language-Action Model

Jinhui Ye, Fangjing Wang, Ning Gao, Junqiu Yu, Zhu Yangkun, Bin Wang, Jinyu Zhang, Weiyang Jin, Yanwei Fu, Feng Zheng, Yilun Chen, Jiangmiao Pang

ICLR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Vision-language-action foundation models

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 spatial grounding pretraining and spatial promptingVanilla VLA (Qwen2.5-VL-3B) · SimplerEnv Google Robot VMsuccess rate (%): 66.1 -> 84.6Improved
add spatial grounding pretraining and spatial promptingVanilla VLA · SimplerEnv WidowX VMsuccess rate (%): 54.7 -> 73.2Improved
replace vanilla co-training with spatially guided training (ablation)VLA · SimplerEnv WidowX VMsuccess rate (%): 61.1 -> 73.2Improved
add spatial prompting during co-training (ablation)VLA co-training · gradient alignment PSSPSS: 0.25 -> 0.42Improved
add spatial pretraining with robotic grounding data (vs general grounding data) (ablation)ST4VLA · Google Robot VMsuccess rate (%): 72.6 -> 84.3Improved
use general grounding data pretraining instead of none (ablation)VLA · WidowX VMsuccess rate (%): 54.9 -> 65.2Improved
raise grounding vs action co-train ratio from 1:1 to 1:10 (ablation)ST4VLA · Google Robot VMsuccess rate (%): 52.4 -> 80.7Improved
raise co-train ratio beyond 1:10 to 1:20 (ablation)ST4VLA · Google Robot VMsuccess rate (%): 80.7 -> 78.3Hurt
scale spatial pretraining data from 0M to 3M pairs (ablation)ST4VLA · SimplerEnv averagesuccess rate (%): 61.4 -> 77.9Improved
use 0.5M pairs of spatial pretraining (ablation)ST4VLA · SimplerEnv averagesuccess rate (%): 61.4 -> 61.0No clear effect
replace unified prompt with random padding (ablation)ST4VLA · SimplerEnv averagesuccess rate (%): 77.9 -> 58.5Hurt
replace unified prompt with box/point/trace constrained prompt (ablation)ST4VLA · SimplerEnv averagesuccess rate (%): 77.9 -> 76.6 (box), 74.9 (point), 73.9 (trace)Hurt

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