Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Curriculum Learning for 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 |
|---|---|---|---|
| Train with self-paced curriculum learning instead of uniform mini-batch training | Follower · R2R validation unseen | SR (%): 28.6 -> 32.2 | Helped |
| Train with self-paced curriculum learning instead of uniform mini-batch training | Self-Monitoring · R2R validation unseen | SR (%): 38.4 -> 41.0 | Helped |
| Train with self-paced curriculum learning instead of uniform mini-batch training | EnvDrop · R2R validation unseen | SR (%): 45.7 -> 47.6 | Helped |
| Train with self-paced curriculum learning instead of uniform mini-batch training | Follower · R2R validation seen | SR (%): 52.3 -> 58.7 | Helped |
| Train with self-paced curriculum learning instead of uniform mini-batch training | Self-Monitoring · R2R validation seen | SR (%): 58.4 -> 58.8 | No clear effect |
| Train with self-paced curriculum learning instead of uniform mini-batch training | EnvDrop · R2R validation seen | SR (%): 57.7 -> 58.1 | No clear effect |
| Present difficulty-organized samples sequentially from easy to hard | Follower · R2R validation unseen | SR (%): 28.6 -> 31.1 | Helped |
| Present difficulty-organized samples sequentially from easy to hard | Follower · R2R validation seen | SR (%): 52.3 -> 48.6 | Hurt |
| Present difficulty-organized samples sequentially from easy to hard | Self-Monitoring · R2R validation unseen | SR (%): 38.4 -> 40.0 | Helped |
| Randomize the order of difficulty-organized curriculum splits (ablation) | Self-Monitoring with naïve curriculum learning · R2R validation unseen | SR (%): 38.4 -> 36.4 | Hurt |
| Feed curriculum splits from hardest to easiest (ablation) | Self-Monitoring with naïve curriculum learning · R2R validation unseen | SR (%): 38.4 -> 38.8 | No clear effect |
| Train first on R2R, then transfer to RxR-en | Self-Monitoring · RxR-en validation unseen | SR (%): 16.0 -> 22.1 | Helped |
Closest papers by meaning
- Towards Learning a Generic Agent for Vision-and-Language Navigation via Pre-Training (CVPR 2020 · similarity 0.92)
- A New Path: Scaling Vision-and-Language Navigation With Synthetic Instructions and Imitation Learning (CVPR 2023 · similarity 0.91)
- Behavioral Analysis of Vision-and-Language Navigation Agents (CVPR 2023 · similarity 0.91)
- Human-Aware Vision-and-Language Navigation: Bridging Simulation to Reality with Dynamic Human Interactions (NeurIPS 2024 · similarity 0.91)
- General Scene Adaptation for Vision-and-Language Navigation (ICLR 2025 · similarity 0.89)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · 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.