Research map / Vision-language and multimodal / Robot action and navigation
Chasing Ghosts: Instruction Following as Bayesian State Tracking
Peter Anderson, Ayush Shrivastava, Devi Parikh, Dhruv Batra, Stefan Lee
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 |
|---|---|---|---|
| replace single-step LingUNet with Bayes filter over trajectories (motion and observation models) | goal prediction on semantic map · R2R val-unseen | success rate (<3m error, %): 32.2 -> 44.9 | Improved |
| replace LingUNet with Bayes filter | goal prediction on semantic map · R2R val-seen | prediction error (m): 5.74 -> 5.09 | Improved |
| add heading theta to filter state (ablation) | Bayes filter · R2R val-unseen | success rate (<3m error, %): 39.4 -> 44.9 | Improved |
| add heading theta to filter state (ablation) | Bayes filter · R2R val-seen | prediction error (m): 5.31 -> 5.09 | Improved |
| Bayes filter vs hand-coded dataset-bias baseline | goal prediction · R2R val-unseen | success rate (<3m error, %): 21.4 -> 44.9 | Improved |
| full VLN agent without RL or augmentation versus RL/augmentation agents | full VLN with reactive policy · R2R test | success rate: 0.51 -> 0.33 | Hurt |
| compare against methods without augmentation (Speaker-Follower) | full VLN · R2R val-unseen | success rate: 0.31 -> 0.35 | Mixed |
Closest papers by meaning
- Learning to Map for Active Semantic Goal Navigation (ICLR 2022 · similarity 0.89)
- Progress-Think: Semantic Progress Reasoning for Vision-Language Navigation (CVPR 2026 · similarity 0.89)
- CompassNav: Steering From Path Imitation to Decision Understanding In Navigation (ICLR 2026 · similarity 0.89)
- Trajectory Diffusion for ObjectGoal Navigation (NeurIPS 2024 · similarity 0.88)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · similarity 0.87)
- Scene Map-based Prompt Tuning for Navigation Instruction Generation (CVPR 2025 · 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.
