Research map / Reinforcement learning / Visual and representation RL
Learning Active Camera for Multi-Object Navigation
Peihao Chen, Dongyu Ji, Kunyang Lin, Weiwen Hu, Wenbing Huang, Thomas H. Li, Mingkui Tan, Chuang Gan
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 EXPO camera policy | OccAnt · MatterPort3D 3-ON | SPL: 53.0 $\pm 0.0$ -> 57.9 $\pm 0.5$ | Improved |
| Add EXPO camera policy | OccAnt · Gibson 3-ON | SPL: 76.1 $\pm 0.1$ -> 78.9 $\pm 0.5$ | Improved |
| Add EXPO camera policy | Mapping + FBE · MatterPort3D 3-ON | SPL: 40.1 $\pm 0.0$ -> 44.6 $\pm 0.3$ | Improved |
| Add EXPO camera policy | Mapping + FBE · Gibson 3-ON | SPL: 62.1 $\pm 0.3$ -> 68.7 $\pm 0.1$ | Improved |
| Add EXPO camera policy | MultiON · MatterPort3D 3-ON | SPL: 33.0 $\pm 0.5$ -> 38.7 $\pm 0.7$ | Improved |
| Add EXPO camera policy | DD-PPO · MatterPort3D 3-ON | SPL: 16.7 $\pm 0.2$ -> 19.1 $\pm 0.4$ | Improved |
| Add random camera actions | OccAnt · MatterPort3D 3-ON | SPL: 53.0 $\pm 0.0$ -> 48.7 $\pm 0.5$ | Hurt |
| Add random camera actions | MultiON · MatterPort3D 3-ON | SPL: 33.0 $\pm 0.5$ -> 3.1 $\pm 0.6$ | Hurt |
| Add occupancy-map input to EXPO (ablation) | OccAnt + EXPO camera policy · MatterPort3D 3-ON | SPL: 55.9 $\pm 0.7$ -> 57.9 $\pm 0.5$ | Improved |
| Add heuristic-direction input to EXPO (ablation) | OccAnt + EXPO camera policy · MatterPort3D 3-ON | SPL: 56.5 $\pm 0.6$ -> 57.9 $\pm 0.5$ | Improved |
| Add navigation-action input to EXPO (ablation) | OccAnt + EXPO camera policy · MatterPort3D 3-ON | SPL: 56.9 $\pm 0.4$ -> 57.9 $\pm 0.5$ | Improved |
| Add heuristic reward to EXPO (ablation) | OccAnt + EXPO camera policy · MatterPort3D 3-ON | SPL: 53.8 $\pm 0.5$ -> 57.9 $\pm 0.5$ | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DD-PPO — DD-PPO: Learning Near-Perfect PointGoal Navigators from 2.5 Billion Frames (ICLR 2020) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Learning on the Go: A Meta-learning Object Navigation Model (ICCV 2025 · similarity 0.88)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 2025 · similarity 0.88)
- ACTIVE-o3 : Empowering MLLMs with Active Perception via Pure Reinforcement Learning (ICML 2026 · similarity 0.87)
- Learning with a Mole: Transferable latent spatial representations for navigation without reconstruction (ICLR 2024 · similarity 0.86)
- Active Test-time Vision-Language Navigation (NeurIPS 2025 · similarity 0.86)
- Navigation World Models (CVPR 2025 · 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.
