Research map / Vision-language and multimodal / Robot action and navigation
KERM: Knowledge Enhanced Reasoning for Vision-and-Language Navigation
Read the paperCheck 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 |
|---|---|---|---|
| Add retrieved knowledge facts and KERM reasoning to DUET | DUET · REVERIE Val Unseen | SR: 46.98 -> 50.44 | Helped |
| Add retrieved knowledge facts and KERM reasoning to DUET | DUET · R2R Val Unseen | SR: 71.52 -> 71.95 | Helped |
| Add retrieved knowledge facts and KERM reasoning to DUET | DUET · SOON Val Unseen | SR: 36.28 -> 38.05 | Helped |
| Add purification (ablation) | KERM with both interaction modules · REVERIE Val Unseen | SR: 49.43 -> 50.44 | Helped |
| Add vision-fact interaction (ablation) | KERM with purification and history-fact interaction · REVERIE Val Unseen | SR: 49.33 -> 50.44 | Helped |
| Add history-fact interaction (ablation) | KERM with purification and vision-fact interaction · REVERIE Val Unseen | SR: 48.96 -> 50.44 | Helped |
| Replace object-centric features with retrieved facts (ablation) | Baseline on REVERIE · REVERIE Val Unseen | SR: 49.73 -> 50.44 | Helped |
| Add object-centric features (ablation) | Baseline on REVERIE · REVERIE Val Unseen | SR: 46.98 -> 49.73 | Helped |
| Use 5 cropped sub-regions instead of 1 (ablation) | KERM · REVERIE Val Unseen | SR: 49.67 -> 50.44 | Helped |
| Use 9 cropped sub-regions instead of 5 (ablation) | KERM · REVERIE Val Unseen | SR: 50.44 -> 48.99 | Hurt |
| Use KERM pretraining with augmented data before fine-tuning (ablation) | KERM · REVERIE Val Unseen | SR: 49.02 -> 50.44 | Helped |
| Use knowledge-enhanced features in the fine-scale encoder (ablation) | DUET-fine · REVERIE Val Unseen | SR: 28.86 -> 30.80 | Helped |
Closest papers by meaning
- AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation (CVPR 2026 · similarity 0.91)
- Aux-Think: Exploring Reasoning Strategies for Data-Efficient Vision-Language Navigation (NeurIPS 2025 · similarity 0.90)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · similarity 0.89)
- FantasyVLN: Unified Multimodal Chain-of-Thought Reasoning for Vision-and-Language Navigation (CVPR 2026 · similarity 0.89)
- Augmented Commonsense Knowledge for Remote Object Grounding (AAAI 2024 · similarity 0.89)
- SenseSearch: Empowering Vision-Language Models with High-Resolution Agentic Search-Reasoning via Reinforcement Learning (CVPR 2026 · similarity 0.87)
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.