Research map / Multi-agent systems and games / Human-AI and goal-conditioned
Learning to Move with Affordance Maps
William Qi, Ravi Teja Mullapudi, Saurabh Gupta, Deva Ramanan
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 learned affordance maps to frontier-based exploration | Frontier-based exploration · Hazard-dense environments | exploration performance: approximately 60% | Improved |
| Add learned affordance maps to frontier-based exploration | Frontier-based exploration · Hazard-sparse environments | exploration performance | No clear effect |
| Use affordance-augmented frontier exploration instead of PPO-based exploration | Exploration methods · Hazard-dense environments | exploration performance: 70% | Improved |
| Add affordance maps to geometry-based A* navigation | Geometry-based navigation · Goal-directed navigation | success rate: 45% | Improved |
| Train the affordance model on a smaller dataset, one fifth the size (ablation) | Affordance-augmented navigation · Goal-directed navigation | success rate: 25% | Improved |
| Collect training data using actively planned trajectories instead of random samples alone (ablation) | Affordance-augmented navigation · Goal-directed navigation | success rate: more than 10% | Improved |
| Collect active training data rather than relying on random samples alone (ablation) | Affordance-augmented navigation · Goal-directed navigation | success rate: using just 40% of the data | Improved |
| Increase geometry-baseline replanning frequency 10-fold (ablation) | Geometry-based navigation · Goal-directed navigation | success rate: 10-fold | Improved |
Closest papers by meaning
- An Interactive Navigation Method with Effect-oriented Affordance (CVPR 2024 · similarity 0.91)
- Learning with a Mole: Transferable latent spatial representations for navigation without reconstruction (ICLR 2024 · similarity 0.89)
- GAMap: Zero-Shot Object Goal Navigation with Multi-Scale Geometric-Affordance Guidance (NeurIPS 2024 · similarity 0.88)
- Imagine Before Go: Self-Supervised Generative Map for Object Goal Navigation (CVPR 2024 · similarity 0.88)
- Affordances-Oriented Planning Using Foundation Models for Continuous Vision-Language Navigation (AAAI 2025 · similarity 0.87)
- AffordDP: Generalizable Diffusion Policy with Transferable Affordance (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.
