Research map / Reinforcement learning / Visual and representation RL
Object-Centric World Models for Causality-Aware Reinforcement Learning
Yosuke Nishimoto, Takashi Matsubara
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 object-centric world model alone (no object-centric policy/value) (ablation) | STICA · Safety Gym PointButton1, CarButton1 | return | No clear effect |
| add Transformer policy and value networks over object tokens (ablation) | STICA · Safety Gym PointButton1, CarButton1 | return | Improved |
| add background removal via static background latent (ablation) | STICA · Safety Gym PointButton1, CarButton1 | return | Improved |
| add causal attention in policy and value networks (ablation) | STICA · Safety Gym PointButton1, CarButton1 | return | Improved |
| use STICA instead of DreamerV3 | STICA · Safety Gym PointButton1 | return: 4.01 -> 11.52 | Improved |
| use STICA instead of DreamerV3 | STICA · Safety Gym PointGoal1 | return: 19.13 -> 13.63 | Hurt |
| use STICA instead of TWM | STICA · OCVRL Object Interaction | success rate: 0.080 -> 0.333 | Improved |
| use STICA instead of best baseline | STICA · Safety Gym (normalized mean) | normalized mean return: 4.06 -> 5.49 | Improved |
| use STICA instead of DreamerV3 | STICA · Safety Gym PointGoal2 | return: 13.64 -> 13.64 | No clear effect |
Closest papers by meaning
- Learning Interactive World Model for Object-Centric Reinforcement Learning (NeurIPS 2025 · similarity 0.88)
- Efficient Exploration and Discriminative World Model Learning with an Object-Centric Abstraction (ICLR 2025 · similarity 0.87)
- Modeling Unseen Environments with Language-guided Composable Causal Components in Reinforcement Learning (ICLR 2025 · similarity 0.87)
- SOLD: Slot Object-Centric Latent Dynamics Models for Relational Manipulation Learning from Pixels (ICML 2025 · similarity 0.87)
- Beyond Prediction: Steering VLM Agents with Retrospective World Modeling (NeurIPS 2026 · similarity 0.86)
- Learning Causal Dynamics Models in Object-Oriented Environments (ICML 2024 · 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.
