Research map / Vision-language and multimodal / Robot action and navigation
Waypoint Models for Instruction-Guided Navigation in Continuous Environments
Jacob Krantz, Aaron Gokaslan, Dhruv Batra, Stefan Lee, Oleksandr Maksymets
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 |
|---|---|---|---|
| use continuous instead of discrete offset prediction (ablation) | WPN · VLN-CE Val-Unseen | success rate: 28 -> 36 | Improved |
| use continuous offset instead of fixed zero offset (ablation) | HPN · VLN-CE Val-Unseen | success rate: 34 -> 38 | Improved |
| use fixed zero offset instead of dense discrete offsets (ablation) | WPN · VLN-CE Val-Unseen | success rate: 28 -> 31 | Improved |
| use continuous instead of discrete distance prediction (ablation) | WPN · VLN-CE Val-Unseen | success rate: 36 -> 34 | No clear effect |
| remove distance prediction (fixed 0.25m step), keep continuous offset (ablation) | WPN · VLN-CE Val-Unseen | success rate: 36 -> 38 | Improved |
| remove distance prediction (fixed 0.25m step), keep continuous offset (ablation) | WPN · VLN-CE Val-Unseen | SPL: 30 -> 36 | Improved |
| add distance prediction to heading-only model (ablation) | HPN · VLN-CE Val-Unseen | estimated execution time (s): 297 -> 153 | Improved |
| add distance prediction to heading-only model (ablation) | HPN · VLN-CE Val-Unseen | SCT: 11 -> 23 | Improved |
| execute waypoints with discrete navigator instead of continuous navigator | WPN · VLN-CE Val-Unseen | success rate: 36 -> 33 | Hurt |
| use single forward-facing camera instead of panorama (ablation) | HPN pick-pano · VLN-CE Val-Unseen | success rate: 34 -> 36 | Improved |
| train with slack reward -0.05 instead of -0.01 (ablation) | WPN · VLN-CE Val-Unseen | SPL | Improved |
| use HPN+DN instead of CMA+PM+DA+Aug | VLN-CE agent · VLN-CE Test | success rate: 28 -> 32 | Improved |
Closest papers by meaning
- Bridging the Gap Between Learning in Discrete and Continuous Environments for Vision-and-Language Navigation (CVPR 2022 · similarity 0.91)
- Learning to Follow and Generate Instructions for Language-Capable Navigation (TPAMI 2023 · similarity 0.89)
- Object-and-Action Aware Model for Visual Language Navigation (ECCV 2020 · similarity 0.89)
- OmniNav: A Unified Framework for Prospective Exploration and Visual-Language Navigation (ICLR 2026 · similarity 0.89)
- STRIDER: Navigation via Instruction-Aligned Structural Decision Space Optimization (NeurIPS 2025 · similarity 0.89)
- Affordances-Oriented Planning Using Foundation Models for Continuous Vision-Language Navigation (AAAI 2025 · similarity 0.88)
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.
