Research map / Vision-language and multimodal / Robot action and navigation
Vision-Language Navigation With Self-Supervised Auxiliary Reasoning Tasks
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 |
|---|---|---|---|
| Add all four auxiliary reasoning tasks (ablation) | VLN baseline · R2R val seen | SPL (%): 55.87 -> 58.89 | Helped |
| Add all four auxiliary reasoning tasks (ablation) | VLN baseline · R2R val unseen | SPL (%): 42.89 -> 45.67 | Helped |
| Add cross-modal matching alone (ablation) | VLN baseline · R2R val seen | SPL (%): 55.87 -> 53.55 | Hurt |
| Add trajectory retelling (ablation) | VLN baseline · R2R val unseen | SPL (%): 42.89 -> 45.24 | Helped |
| Add progress estimation (ablation) | VLN baseline · R2R val unseen | SPL (%): 42.89 -> 44.74 | Helped |
| Add angle prediction (ablation) | VLN baseline · R2R val unseen | SPL (%): 42.89 -> 43.77 | No clear effect |
| Add all four auxiliary tasks alongside back-translation (ablation) | VLN baseline + back-translation · R2R val seen | SPL (%): 61.37 -> 67.17 | Helped |
| Use teacher forcing instead of Matching Critic for trajectory retelling (ablation) | VLN baseline · R2R val unseen | SPL (%): 43.44 -> 45.24 | Helped |
| Use teacher forcing instead of student forcing for trajectory retelling (ablation) | VLN baseline · R2R val unseen | SPL (%): 43.78 -> 45.24 | Helped |
| Share context in teacher-forced trajectory retelling (ablation) | VLN baseline · R2R val unseen | SPL (%): 45.24 -> 44.47 | Hurt |
| Use step-percentage progress labels with BCE rather than MSE loss (ablation) | VLN progress estimator · R2R val seen | SPL (%): 50.81 -> 57.15 | Helped |
| Use step-percentage progress labels with BCE rather than MSE loss (ablation) | VLN progress estimator · R2R val unseen | SPL (%): 43.16 -> 44.74 | Helped |
Closest papers by meaning
- Aux-Think: Exploring Reasoning Strategies for Data-Efficient Vision-Language Navigation (NeurIPS 2025 · similarity 0.92)
- AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation (CVPR 2026 · similarity 0.92)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · similarity 0.90)
- FantasyVLN: Unified Multimodal Chain-of-Thought Reasoning for Vision-and-Language Navigation (CVPR 2026 · similarity 0.89)
- Hierarchical Semantic-Augmented Navigation: Optimal Transport and Graph-Driven Reasoning for Vision-Language Navigation (NeurIPS 2025 · similarity 0.89)
- Towards Long-Horizon Vision-Language Navigation: Platform, Benchmark and Method (CVPR 2025 · similarity 0.89)
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.