Research map / Detection and segmentation / Open-set and few-shot detection
Reasoning About Physical Interactions with Object-Oriented Prediction and Planning
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 |
|---|---|---|---|
| Use learned physics simulation when evaluating candidate placements (ablation) | O2P2 tower-building planner · Block tower building | accuracy (%): 0 -> 76 | Helped |
| Use an object-factorized learned model instead of object-agnostic SAVP | Block tower-building planner · Block tower building | accuracy (%): 24 -> 76 | Helped |
| Use object representations for proposal costs rather than pixel-wise costs | Oracle planner · Block tower building | accuracy (%): 71 -> 92 | Helped |
| Use the learned O2P2 model rather than MuJoCo with pixel-space costs | Block tower-building planner · Block tower building | accuracy (%): 71 -> 76 | Helped |
| Transfer the trained model to real-image Sawyer planning | O2P2 · Sawyer real-image goal configurations | correctly built configurations | Helped |
Closest papers by meaning
- Relational Neural Expectation Maximization: Unsupervised Discovery of Objects and their Interactions (ICLR 2018 · similarity 0.88)
- Visual Grounding of Learned Physical Models (ICML 2020 · similarity 0.87)
- A Compositional Object-Based Approach to Learning Physical Dynamics (ICLR 2017 · similarity 0.87)
- Visual Interaction Networks: Learning a Physics Simulator from Video (NeurIPS 2017 · similarity 0.87)
- Object Concepts Emerge from Motion (NeurIPS 2025 · similarity 0.86)
- OBJECT-UNI: A Unified Model for Object-Centric Spatial Understanding and Controllable Generation (NeurIPS 2026 · similarity 0.85)
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.