Research map / Vision-language and multimodal / Robot action and navigation
From Spatial to Actions: Grounding Vision-Language-Action Model in Spatial Foundation Priors
Zhengshen Zhang, Hao Li, Yalun Dai, Zhengbang Zhu, Lei Zhou, Chenchen Liu, Dong Wang, Francis E. H. Tay, Sijin Chen, Ziwei Liu, Yuxiao Liu, Xinghang Li and 1 more
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 changed | Where | Measured | Effect |
|---|---|---|---|
| inject spatial tokens into the action head instead of the VLM (ablation) | FALCON · CALVIN ABC->D | Avg. Len.: 3.79 -> 3.91 | Improved |
| inject spatial tokens into the action head instead of the VLM (ablation) | FALCON · CALVIN ABCD->D | Avg. Len.: 4.00 -> 4.08 | Improved |
| use element-wise addition instead of cross-attention fusion (ablation) | FALCON · CALVIN ABC->D | Avg. Len.: 3.68 -> 3.91 | Improved |
| use element-wise addition instead of FiLM-gated fusion (ablation) | FALCON · CALVIN ABC->D | Avg. Len.: 3.76 -> 3.91 | Improved |
| add test-time depth to RGB-trained FALCON | FALCON · CALVIN ABC->D | Avg. Len.: 3.91 -> 3.95 | Improved |
| add test-time depth to RGB-trained FALCON | FALCON · CALVIN ABCD->D | Avg. Len.: 4.08 -> 4.09 | No clear effect |
| remove depth at test time for RGB-D trained FALCON | FALCON · CALVIN ABCD->D | Avg. Len.: 4.09 -> 4.07 | No clear effect |
| replace point cloud encoder with ESM (RGB-only) over Kosmos-VLA RGB | FALCON · CALVIN ABC->D | Avg. Len.: 3.48 -> 3.91 | Improved |
| add wrist camera images to ESM (ablation) | FALCON · CALVIN ABCD->D | Avg. Len.: 4.08 -> 4.10 | Improved |
| increase loss weight lambda from 0.01 to 0.05 (ablation) | FALCON · CALVIN ABC->D | Avg. Len.: 3.91 -> 3.87 | Hurt |
| increase LSTM history length from 8 to 16 (ablation) | Kosmos-VLA (LSTM) · CALVIN ABC->D | Avg. Len.: 3.90 -> 4.23 | Improved |
| increase LSTM chunk size from 5 to 10 (ablation) | Kosmos-VLA (LSTM) · CALVIN ABC->D | Avg. Len.: 3.99 -> 4.23 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- UP-VLA — UP-VLA: A Unified Understanding and Prediction Model for Embodied Agent (ICML 2025) · better on 2 of 2 shared results
- TraceVLA — TraceVLA: Visual Trace Prompting Enhances Spatial-Temporal Awareness for Generalist Robotic Policies (ICLR 2025) · better on 1 of 1 shared result
- Seer-Large — SEER: Backdoor Detection for Vision-Language Models through Searching Target Text and Image Trigger Jointly (AAAI 2024) · better on 1 of 1 shared result
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
- SpatialRGPT: Grounded Spatial Reasoning in Vision-Language Models (NeurIPS 2024 · similarity 0.90)
- SpaceMind: Camera-Guided Modality Fusion for Spatial Reasoning in Vision-Language Models (CVPR 2026 · similarity 0.90)
- From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D (NeurIPS 2025 · similarity 0.89)
- Abstract 3D Perception for Spatial Intelligence in Vision-Language Models (CVPR 2026 · similarity 0.89)
- Grounded 3D-Aware Spatial Vision-Language Modeling (CVPR 2026 · similarity 0.89)
- ConsisVLA-4D: Advancing Spatiotemporal Consistency in Efficient 3D-Perception and 4D-Reasoning for Robotic Manipulation (CVPR 2026 · similarity 0.89)
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.
