Research map / Vision-language and multimodal / Robot action and navigation
ASSISTER: Assistive Navigation via Conditional Instruction Generation
Zanming Huang, Zhongkai Shangguan, Jimuyang Zhang, Gilad Bar, Matthew Boyd, Eshed Ohn-Bar
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 |
|---|---|---|---|
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Sim (training conditions) | CIDEr: 16.97 -> 23.58 | Improved |
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Sim (new town) | CIDEr: 14.49 -> 19.93 | Improved |
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Sim (new town and weather) | CIDEr: 11.16 -> 19.61 | Improved |
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Sim | SPICE: 23.98 -> 26.31 (training), 18.69 -> 18.63 (new town) | Mixed |
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Real (cross-subject) | CIDEr: 10.42 -> 10.74 | No clear effect |
| add command-conditional module to goal-only model (ablation) | ASSISTER · UrbanWalk-Sim (training conditions) | BLEU-4: 10.58 -> 15.49 | Improved |
| test on new town instead of training conditions (ablation) | ASSISTER (Goal+Command) · UrbanWalk-Sim | CIDEr: 23.58 -> 19.93 | Hurt |
| test on new town and unseen weather instead of new town only (ablation) | ASSISTER (Goal+Command) · UrbanWalk-Sim | CIDEr: 19.93 -> 19.61 | Hurt |
| evaluate on real-world data instead of simulation | ASSISTER (Goal+Command) · UrbanWalk | CIDEr: 23.58 -> 10.74 | Hurt |
| use ASSISTER instructions instead of Xu et al. baseline | human followers in blind simulation · Task-based evaluation | success rate: 9.52 -> 38.1 | Improved |
| use ASSISTER instructions instead of Xu et al. baseline | human followers in blind simulation · Task-based evaluation | route completion: 46.3 -> 74.1 | Improved |
| use ASSISTER instructions instead of Xu et al. baseline | human followers in blind simulation · Task-based evaluation | collision with static obstacles: 6.0 -> 14.0 | Hurt |
Closest papers by meaning
- Vision-Based Navigation With Language-Based Assistance via Imitation Learning With Indirect Intervention (CVPR 2019 · similarity 0.89)
- Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-Training (CVPR 2020 · similarity 0.87)
- Learning to Follow and Generate Instructions for Language-Capable Navigation (TPAMI 2023 · similarity 0.87)
- A New Path: Scaling Vision-and-Language Navigation With Synthetic Instructions and Imitation Learning (CVPR 2023 · similarity 0.87)
- Curriculum Learning for Vision-and-Language Navigation (NeurIPS 2021 · similarity 0.86)
- Human-Aware Vision-and-Language Navigation: Bridging Simulation to Reality with Dynamic Human Interactions (NeurIPS 2024 · similarity 0.86)
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.
