Research map / Detection and segmentation / Open-set and few-shot detection
Learning Long-term Visual Dynamics with Region Proposal Interaction Networks
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 RoI features with surrounding scene context instead of global encoding, masked images, or cropped object images | RPIN · PHYRE-W | prediction error: 6.45 -> 1.31 | Helped |
| use RoI features with surrounding scene context instead of global encoding, masked images, or cropped object images | RPIN · RealB | prediction error: 0.59 -> 0.30 | Helped |
| use RoI features with surrounding scene context instead of global encoding, masked images, or cropped object images | RPIN · SimB | prediction error: 3.48 -> 2.55 | Helped |
| use RoI features with surrounding scene context instead of global encoding, masked images, or cropped object images | RPIN · ShapeStacks | prediction error: 2.47 -> 1.03 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · PHYRE-W | prediction error: 1.50 -> 1.31 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · ShapeStacks | prediction error: 1.85 -> 1.03 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · RealB | prediction error: 0.37 -> 0.30 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · SimB | prediction error: 3.01 -> 2.55 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · PHYRE-C | prediction error: 10.98 -> 9.22 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · SS-4 | prediction error: 15.02 -> 14.61 | Helped |
| replace vector-form interaction reasoning with convolutional interaction reasoning over spatial feature maps (ablation) | RPIN · SimB-5 | prediction error: 24.42 -> 22.38 | Helped |
| use RPIN dynamics features for task scoring rather than DQN scoring | PHYRE planning · PHYRE B-tier, within-task | AUCCESS: 77.6 ± 1.1 -> 85.2 ± 0.7 | Helped |
Closest papers by meaning
- Visual Interaction Networks: Learning a Physics Simulator from Video (NeurIPS 2017 · similarity 0.91)
- A Critical View of Vision-Based Long-Term Dynamics Prediction Under Environment Misalignment (ICML 2023 · similarity 0.90)
- Attention over Learned Object Embeddings Enables Complex Visual Reasoning (NeurIPS 2021 · similarity 0.88)
- SOLD: Slot Object-Centric Latent Dynamics Models for Relational Manipulation Learning from Pixels (ICML 2025 · similarity 0.87)
- PlaySlot: Learning Inverse Latent Dynamics for Controllable Object-Centric Video Prediction and Planning (ICML 2025 · similarity 0.86)
- Trimming the Long-Tail of Visual World Modeling Evaluation (NeurIPS 2026 · 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.