An asynchronous architecture decouples incremental voxel-based mapping from VLM-based semantic enrichment to produce queryable open-vocabulary 3D scene graphs that match or exceed prior methods on segmentation and grounding benchmarks.
Chang, Manolis Savva, Maciej Halber, Thomas Funkhouser, and Matthias Niessner
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
UNVERDICTED 6roles
background 1polarities
background 1representative citing papers
Uses VLMs to detect instance concepts and LLMs to infer abstract relationships, assembling them into 3D scene graph forests that are evaluated on uHumans2 and ScanNet and tested in open-vocabulary retrieval on a Spot robot.
Introduces LSM that outputs calibrated multimodal spatial distributions from language plus scene graph, fused via VL-Map to improve 3D target localization on VLA-3D benchmark and real robot.
Frontier VLMs show basic single-step view-action knowledge but fail at multi-turn composition in 3D; an iterative self-exploration and view-graph-distillation framework lifts Qwen2.5-VL-7B to 47.8% success, beating larger models.
Embodied-R1.5 is an 8B EFM achieving SOTA on 16 of 24 embodied VLM benchmarks, fine-tunable to outperform leading VLAs, with claimed zero-shot real-robot generalization.
The survey frames VLA models as pipelines that generate progressively grounded action tokens and classifies those tokens into eight types to guide future development.
citing papers explorer
-
A Survey on Vision-Language-Action Models: An Action Tokenization Perspective
The survey frames VLA models as pipelines that generate progressively grounded action tokens and classifies those tokens into eight types to guide future development.