Research map / Multi-agent systems and games / Human-AI and goal-conditioned
An Exploration of Embodied Visual Exploration
Santhosh Kumar Ramakrishnan, Dinesh Jayaraman, Kristen Lorraine Grauman
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 |
|---|---|---|---|
| use smooth-coverage reward instead of area-coverage (ablation) | coverage exploration · AVD | area visited: 0.85 -> 0.88 | Improved |
| use smooth-coverage reward instead of area-coverage (ablation) | coverage exploration · AVD | view localization: 0.43 -> 0.41 | Hurt |
| use smooth-coverage reward instead of area-coverage (ablation) | coverage exploration · MP3D | area visited: 0.54 -> 0.58 | Improved |
| use smooth-coverage reward instead of area-coverage (ablation) | coverage exploration · MP3D | objects visited: 0.52 -> 0.56 | Improved |
| use smooth-coverage reward instead of area-coverage (ablation) | coverage exploration · MP3D | reconstruction: 0.44 -> 0.44 | No clear effect |
| use RNN spatio-temporal memory features instead of pretrained CNN image features (ablation) | curiosity exploration · MP3D | visitation metrics | Improved |
| train with noisy odometry (eta=0.15) instead of noise-free (ablation) | learned exploration methods · MP3D noisy odometry | exploration performance | Hurt |
| increase number of training environments (10, 20, 40, full) (ablation) | novelty and smooth-coverage · MP3D | exploration performance | Improved |
| use ground-truth occupancy maps instead of inferred occupancy (ablation) | frontier-exploration · MP3D | area visited: 0.579 -> 0.748 | Improved |
| use imitation of oracle trajectories instead of random-actions or forward-action baselines | imitation · AVD and MP3D | visitation and tasks | Improved |
Closest papers by meaning
- Habitat: A Platform for Embodied AI Research (ICCV 2019 · similarity 0.91)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 2025 · similarity 0.89)
- Image2Sim: Scaling Embodied Navigation via Generative Neural Simulator (NeurIPS 2026 · similarity 0.89)
- EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI (CVPR 2024 · similarity 0.89)
- Grounding Video Models to Actions through Goal Conditioned Exploration (ICLR 2025 · similarity 0.88)
- 3D-Mem: 3D Scene Memory for Embodied Exploration and Reasoning (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.
