{"slug": "spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of", "title": "SpaR3D-MoE: Adaptive 3D Spatial Reasoning from Sparse Views Meets Geometry-Inductive Mixture-of-Experts", "summary": "Researchers introduced SpaR3D-MoE, a framework that enhances multimodal large language models with adaptive 3D spatial reasoning from sparse RGB inputs. It uses a spatiotemporal manifold sampling mechanism and a geometry-inductive mixture-of-experts to improve performance on spatial tasks, achieving state-of-the-art results on benchmarks like VSI-Bench, ScanQA, and SQA3D.", "body_md": "arXiv:2607.06620v1 Announce Type: new\nAbstract: Recent Multimodal Large Language Models (MLLMs) struggle to bridge the representational gap between 2D semantic understanding and 3D spatial geometry. Existing 3D-aware models either rely on costly 3D-specific data or utilize RGB-only inputs with heuristic sampling and monolithic, shallow fusion, which respectively disrupt essential spatiotemporal connectivity and induce modality contention across diverse spatial tasks. To overcome these bottlenecks, we introduce SpaR3D-MoE, an end-to-end framework that enables adaptive spatial reasoning by equipping MLLMs with geometry-aware capabilities from only sparse RGB inputs. First, we propose an adaptive spatiotemporal manifold sampling mechanism that constructs a geometry-aware spatiotemporal graph to extract informative keyframes, effectively mitigating sequence redundancy while preserving the scene's topological connectivity. Second, we introduce the heterogeneous geometry-inductive Mixture-of-Experts driven by an instruction-pose aware router, which adaptively routes multimodal tokens to specialized experts, resolving the cross-modal contention inherent in monolithic fusion. Extensive experiments on VSI-Bench, ScanQA, and SQA3D demonstrate that our method achieves state-of-the-art performance. Notably, SpaR3D-MoE achieves the highest average score of 63.5 on VSI-Bench, outperforming the strongest baseline by 7.8 absolute points, alongside relative improvements of 35.4% and 51.4% in Route Plan and Relative Direction tasks, respectively.", "url": "https://wpnews.pro/news/spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of", "canonical_source": "https://arxiv.org/abs/2607.06620", "published_at": "2026-07-09 04:00:00+00:00", "updated_at": "2026-07-09 04:28:26.347869+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "computer-vision", "ai-research"], "entities": ["SpaR3D-MoE", "VSI-Bench", "ScanQA", "SQA3D"], "alternates": {"html": "https://wpnews.pro/news/spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of", "markdown": "https://wpnews.pro/news/spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of.md", "text": "https://wpnews.pro/news/spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of.txt", "jsonld": "https://wpnews.pro/news/spar3d-moe-adaptive-3d-spatial-reasoning-from-sparse-views-meets-geometry-of.jsonld"}}