Research map / Vision-language and multimodal / Multimodal LLMs
Move to Understand a 3D Scene: Bridging Visual Grounding and Exploration for Efficient and Versatile Embodied Navigation
Ziyu Zhu, Xilin Wang, Yixuan Li, Zhuofan Zhang, Xiaojian Ma, Yixin Chen, Baoxiong Jia, Wei Liang, Qian Yu, Zhidong Deng, Siyuan Huang, Qing Li
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 vision-language-exploration pre-training (ablation) | MTU3D · HM3D-OVON | SR: 27.8 -> 33.3 | Improved |
| add VLE pre-training (ablation) | MTU3D · GOAT-Bench | SR: 22.2 -> 36.1 | Improved |
| add VLE pre-training (ablation) | MTU3D · SG3D | SR: 22.9 -> 27.9 | Improved |
| keep spatial memory bank across sub-episodes (ablation) | MTU3D · GOAT-Bench Object goals | SR: 10.5 -> 52.6 | Improved |
| keep spatial memory bank (ablation) | MTU3D · GOAT-Bench Description goals | SR: 28.6 -> 71.4 | Improved |
| keep spatial memory bank (ablation) | MTU3D · GOAT-Bench Image goals | SR: 26.7 -> 60.0 | Improved |
| semantic goal-guided exploration instead of nearest-frontier exploration (ablation) | MTU3D · exploration at step 6 | SR: 33.3 -> 50.0 | Improved |
| use MTU3D exploration trajectories for GPT-4V on A-EQA | GPT-4V · A-EQA | LLM-SPL: 7.5 -> 37.0 | Improved |
| MTU3D vs video-based Uni-NaVid on HM3D-OVON path efficiency | MTU3D · HM3D-OVON Val Unseen | SPL: 12.1 vs 19.8 | Hurt |
| train exploration only with optimal frontiers vs randomized/hybrid selection | MTU3D trajectory data · trajectory collection | overfitting | Improved |
Closest papers by meaning
- D3D-VLP: Dynamic 3D Vision-Language-Planning Model for Embodied Grounding and Navigation (CVPR 2026 · similarity 0.93)
- Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation (NeurIPS 2025 · similarity 0.91)
- 3D-Mem: 3D Scene Memory for Embodied Exploration and Reasoning (CVPR 2025 · similarity 0.91)
- Image2Sim: Scaling Embodied Navigation via Generative Neural Simulator (NeurIPS 2026 · similarity 0.91)
- EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI (CVPR 2024 · similarity 0.91)
- Abstract 3D Perception for Spatial Intelligence in Vision-Language Models (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.
