Research map / Vision-language and multimodal / Robot action and navigation
Vision-Based Navigation With Language-Based Assistance via Imitation Learning With Indirect Intervention
Khanh Nguyen, Debadeepta Dey, Chris Brockett, Bill Dolan
Check 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 |
|---|---|---|---|
| Learn a policy that requests language subgoal assistance when needed | VNLA agent · AskNav Test seen | success rate (%): 42.98 -> 52.09 | Improved |
| Learn a policy that requests language subgoal assistance when needed | VNLA agent · AskNav Test unseen | success rate (%): 25.05 -> 34.50 | Improved |
| Replace never requesting help with continuous requests from the beginning | VNLA agent · AskNav Test seen | success rate (%): 28.39 -> 40.33 | Improved |
| Replace never requesting help with continuous requests from the beginning | VNLA agent · AskNav Test unseen | success rate (%): 6.36 -> 20.00 | Improved |
| Use learned help requests instead of random requests | VNLA agent · AskNav Test seen | success rate (%): 42.98 -> 52.09 | Improved |
| Use learned help requests instead of random requests | VNLA agent · AskNav Test unseen | success rate (%): 25.05 -> 34.50 | Improved |
| Use the learned help-request policy instead of the help-requesting teacher | VNLA agent · AskNav Test seen | success rate (%): 52.26 -> 52.09 | No clear effect |
| Add language subgoals to direct interventions | Direct-advisor VNLA agent · AskNav Test unseen | success rate (%): 32.19 -> 34.56 | Improved |
| Replace direct decision overwriting with indirect subgoal interventions | VNLA agent · AskNav Test seen | success rate (%): 52.09 -> 52.09 | No clear effect |
| Replace direct decision overwriting with indirect subgoal interventions | VNLA agent · AskNav Test unseen | success rate (%): 34.56 -> 34.50 | No clear effect |
| Train with room-type goals, then evaluate object finding without room types | VNLA agent with learned help requests · NoRoom Test seen | success rate (%): 43.69 -> 53.85 | Improved |
| Train with room-type goals, then evaluate object finding without room types | VNLA agent with learned help requests · NoRoom Test unseen | success rate (%): 33.41 -> 41.63 | Improved |
Closest papers by meaning
- A New Path: Scaling Vision-and-Language Navigation With Synthetic Instructions and Imitation Learning (CVPR 2023 · similarity 0.91)
- Vision-Language Navigation Policy Learning and Adaptation (TPAMI 2020 · similarity 0.91)
- Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-Training (CVPR 2020 · similarity 0.91)
- Reinforced Cross-Modal Matching and Self-Supervised Imitation Learning for Vision-Language Navigation (CVPR 2019 · similarity 0.90)
- Human-Aware Vision-and-Language Navigation: Bridging Simulation to Reality with Dynamic Human Interactions (NeurIPS 2024 · similarity 0.88)
- Do Visual Imaginations Improve Vision-and-Language Navigation Agents? (CVPR 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.
