Research map / Vision-language and multimodal / Robot action and navigation
Room-and-Object Aware Knowledge Reasoning for Remote Embodied Referring Expression
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| replace REVERIE [37]+FAST with CKR | REVERIE navigation and grounding · REVERIE val-unseen | SPL: 64.67% | Helped |
| replace REVERIE [37]+FAST with CKR | REVERIE navigation and grounding · REVERIE val-unseen | RSR: 46.05% | Helped |
| add ROAA to the transformer base (ablation) | CKR Base-Net · REVERIE val-unseen | SPL: 7.30% -> 8.11% | Helped |
| add room-entity reasoning (ablation) | CKR with ROAA · REVERIE val-unseen | SR: 17.92% -> 18.26% | Helped |
| add object-entity reasoning (ablation) | CKR with ROAA and room-entity reasoning · REVERIE val-unseen | SR: 18.26% -> 19.11% | Helped |
| add external KG reasoning with top-K=5 (ablation) | CKR object-entity reasoning · REVERIE val-unseen | SPL: 6.77% -> 7.51% | Helped |
| increase external KG reasoning to 3 steps (ablation) | CKR object-entity reasoning · REVERIE val-unseen | SPL: 7.51% -> 8.76% | Helped |
| add multi-step internal KG reasoning (ablation) | CKR object-entity reasoning · REVERIE val-unseen | SPL: 8.76% -> 9.80% | Helped |
| set external knowledge capacity to top-K=5 instead of top-K=0 (ablation) | CKR · REVERIE val-unseen | SPL: 8.17% -> 10.68% | Helped |
| increase external knowledge capacity from top-K=5 to top-K=10 (ablation) | CKR · REVERIE val-unseen | SPL: 10.68% -> 10.10% | Hurt |
| add direction-aware loss (ablation) | CKR with ROAA and room- and object-entity reasoning · REVERIE val-unseen | SPL: 8.29% -> 10.56% | Helped |
| add distance-aware search policy (ablation) | CKR with direction-aware loss · REVERIE val-unseen | SPL: 10.56% -> 11.84% | Helped |
Closest papers by meaning
- Augmented Commonsense Knowledge for Remote Object Grounding (AAAI 2024 · similarity 0.89)
- EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents (ICML 2025 · similarity 0.87)
- Embodied Agents Meet Personalization: Investigating Challenges and Solutions Through the Lens of Memory Utilization (ICLR 2026 · similarity 0.86)
- ClawMachine: Learning to Fetch Visual Tokens for Referential Comprehension (ICLR 2025 · similarity 0.86)
- STAR-R1: Multi-View Spatial TrAnsformation Reasoning by Reinforcing Multimodal LLMs (CVPR 2026 · similarity 0.86)
- From Correspondence to Actions: Human-Like Multi-Image Spatial Reasoning in Multi-modal Large Language Models (ICML 2026 · similarity 0.86)
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.