Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Iterative 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 |
|---|---|---|---|
| Extend transformer history across episodes in a tour | HAMT on IR2R · IR2R Val-Unseen | t-nDTW: 50 -> 32 | Hurt |
| Add inflection weighting, a previous-episode identifier, and a trainable extended history encoder (ablation) | TourHAMT on IR2R · IR2R Val-Unseen | t-nDTW: 28 -> 32 | Helped |
| Keep the vision GRU state across episodes (ablation) | CMA on IR2R-CE · IR2R-CE Val-Seen | t-nDTW: 39 -> 40 | Helped |
| Keep the vision GRU state across episodes (ablation) | CMA on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: 38 -> 36 | Hurt |
| Pool episode hidden states into persistent tour memory (ablation) | CMA on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: 38 -> 36 | Hurt |
| Add iterative map construction across episodes during training and evaluation (ablation) | MAP-CMA with ground-truth semantics on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: 42 -> 48 | Helped |
| Add iterative map construction across episodes during training and evaluation (ablation) | MAP-CMA with inferred semantics on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: 43 -> 47 | Helped |
| Use inferred rather than ground-truth semantics with iterative train/eval maps (ablation) | MAP-CMA on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: 48 -> 47 | Hurt |
| Provide known maps rather than iteratively constructed maps at evaluation (ablation) | MAP-CMA with ground-truth semantics and iterative-map training · IR2R-CE Val-Unseen | t-nDTW: 48 -> 49 | Mixed |
| Provide known maps rather than iteratively constructed maps at evaluation (ablation) | MAP-CMA with inferred semantics and iterative-map training · IR2R-CE Val-Unseen | t-nDTW: 47 -> 46 | Hurt |
| Use map-structured memory instead of unstructured CMA memory | MAP-CMA on IR2R-CE · IR2R-CE Val-Unseen | t-nDTW: +9 | Helped |
Closest papers by meaning
- General Scene Adaptation for Vision-and-Language Navigation (ICLR 2025 · similarity 0.91)
- Towards Long-Horizon Vision-Language Navigation: Platform, Benchmark and Method (CVPR 2025 · similarity 0.90)
- Human-Aware Vision-and-Language Navigation: Bridging Simulation to Reality with Dynamic Human Interactions (NeurIPS 2024 · similarity 0.90)
- Hierarchical Semantic-Augmented Navigation: Optimal Transport and Graph-Driven Reasoning for Vision-Language Navigation (NeurIPS 2025 · similarity 0.89)
- STRIDER: Navigation via Instruction-Aligned Structural Decision Space Optimization (NeurIPS 2025 · similarity 0.89)
- NavBench: Probing Multimodal Large Language Models for Embodied Navigation (NeurIPS 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.