Do Embodied VLMs Actually Use Their Scene-Graph Memory?
cs.RO, cs.AI, cs.CV
Submitted: 2026-11-01
Updated: 2026-11-01
Terminology
Sources
- Hydra: A Real-time Spatial Perception System for 3D Scene Graph Construction and Optimization
- SayPlan: Grounding Large Language Models using 3D Scene Graphs for Scalable Robot Task Planning
- Let Your Graph Do the Talking: Encoding Structured Data for LLMs
- LLaGA: Large Language and Graph Assistant
- MomaGraph: State-Aware Unified Scene Graphs with Vision-Language Model for Embodied Task Planning
- Explore until Confident: Efficient Exploration for Embodied Question Answering
- GraphEQA: Using 3D Semantic Scene Graphs for Real-time Embodied Question Answering
- GraphPad: Inference-Time 3D Scene Graph Updates for Embodied Question Answering
- Chat-Scene: Bridging 3D Scene and Large Language Models with Object Identifiers
- GreaseLM: Graph REASoning Enhanced Language Models for Question Answering
- AI2-THOR: An Interactive 3D Environment for Visual AI
- Qwen2.5-VL Technical Report
- Modeling Relational Data with Graph Convolutional Networks
- LoRA: Low-Rank Adaptation of Large Language Models
- Qwen3 Technical Report
- Qwen3-VL Technical Report
- Building and better understanding vision-language models: insights and future directions
Related papers
- Rethinking Actions: Continuity-Aware Priors for World-Model Planning
- MOSAIC: Modular Obstacle and Spatial-Aware Framework for Manipulation In Cluttered Scenes
- SAGAS: Semantic-Aware Graph-Assisted Stitching for Offline Temporal Logic Planning
- DexSeed: Expanding Human Hand-Object Interactions into Robot Demonstrations for Dexterous Policy Learning
- LaST0.5: A Generalist World Action Model with Interaction-Centric Latent Reasoning
- CIALLO: Learning Shared Intent Representations for Cross-Embodiment Manipulation