Research map / Vision-language and multimodal / Multimodal LLMs
SpaR3D-MoE: Adaptive 3D Spatial Reasoning from Sparse Views Meets Geometry-Inductive Mixture-of-Experts
Haida Feng, Hao Wei, Haolin Wang, Shiwei Li, Chade Li, Yihong Wu
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 |
|---|---|---|---|
| add Adaptive Spatiotemporal Manifold Sampling (vs uniform), 32 frames (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 62.4 -> 63.5 | Improved |
| add ASMS (vs uniform), 32 frames (ablation) | SpaR3D-MoE · VSI-Bench Route Plan | accuracy: 39.8 -> 44.0 | Improved |
| add ASMS (vs uniform), 8 frames (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 56.6 -> 57.8 | Improved |
| add ASMS (vs uniform), 16 frames (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 61.2 -> 61.9 | Improved |
| add holistic expert E0 (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 61.8 -> 63.5 | Improved |
| add geometric-semantic cross-attention expert E1 (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 62.8 -> 63.5 | Improved |
| add pose-conditioned dynamic adapter expert E2 (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 59.3 -> 63.5 | Improved |
| add gravity-aligned structural expert E3 (ablation) | SpaR3D-MoE · VSI-Bench | average accuracy: 62.4 -> 63.5 | Improved |
| add camera pose bias to the router (ablation) | Base Router · VSI-Bench | average accuracy: 61.2 -> 62.5 | Improved |
| add instruction query guidance to the router (ablation) | Base Router · VSI-Bench | average accuracy: 61.2 -> 61.8 | Improved |
| combine pose bias and query guidance in the router (ablation) | Base Router · VSI-Bench | average accuracy: 61.2 -> 63.5 | Improved |
| use SpaR3D-MoE instead of Qwen3VL-8B | Qwen3VL-8B · VSI-Bench | average accuracy: 55.7 -> 63.5 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- Spatial-MLLM — Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence (NeurIPS 2025) · better on 3 of 3 shared results
- 3D-LLaVA — 3D-LLaVA: Towards Generalist 3D LMMs with Omni Superpoint Transformer (CVPR 2025) · 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
- SpatialReasoner: Towards Explicit and Generalizable 3D Spatial Reasoning (NeurIPS 2025 · similarity 0.91)
- From Flatland to Space: Teaching Vision-Language Models to Perceive and Reason in 3D (NeurIPS 2025 · similarity 0.91)
- SpatialLLM: A Compound 3D-Informed Design towards Spatially-Intelligent Large Multimodal Models (CVPR 2025 · similarity 0.91)
- Spatial-MLLM: Boosting MLLM Capabilities in Visual-based Spatial Intelligence (NeurIPS 2025 · similarity 0.91)
- SpaceMind: Camera-Guided Modality Fusion for Spatial Reasoning in Vision-Language Models (CVPR 2026 · similarity 0.91)
- SpatialStack: Layered Geometry-Language Fusion for 3D VLM Spatial Reasoning (CVPR 2026 · similarity 0.90)
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.
