Research map / Detection and segmentation / Open-set and few-shot detection
Object-Category Aware Reinforcement Learning
Qi Yi, Rui Zhang, Shaohui Peng, Jiaming Guo, Xing Hu, Zidong Du, Xishan Zhang, Qi Guo, Yunji Chen
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 |
|---|---|---|---|
| replace RRL with OCARL | RRL · Crafter | score: 4.22 -> 12.31 | Improved |
| replace SMORL with OCARL | SMORL · Crafter | score: 3.94 -> 12.31 | Improved |
| disable zombie module f_Z in OCMR (ablation) | OCARL · Hunter 4Z+0C | return: 8.27 -> 1.17 | Hurt |
| disable zombie module f_Z in OCMR (ablation) | OCARL · Hunter 0Z+4C | return: 8.29 -> 8.13 | No clear effect |
| disable cow module f_C in OCMR (ablation) | OCARL · Hunter 0Z+4C | return: 8.29 -> 4.90 | Hurt |
| disable cow module f_C in OCMR (ablation) | OCARL · Hunter 4Z+0C | return: 8.27 -> 8.32 | No clear effect |
| remove OCMR (use universal reasoning network) (ablation) | OCARL · Hunter generalization to 4Z+4C | mean return | Hurt |
| remove OCAP (lambda_cat = 0), keep OCMR (ablation) | OCARL · Hunter generalization to 4Z+4C | mean return | Mixed |
| randomly mask discovered objects with probability p up to 1 (ablation) | OCARL · Hunter, 5M steps | final return | Hurt |
| add object-category supervision and modular reasoning (OCARL) to PPO baselines | PPO · Hunter 4Z+4C generalization from 1Z+1C | mean return | Improved |
| replace RRL with OCARL | RRL · Crafter Collect Stone | success rate (%): 3.2 -> 46.1 | Improved |
| replace SMORL with OCARL | SMORL · Crafter Defeat Zombie | success rate (%): 11.3 -> 51.4 | Improved |
Closest papers by meaning
- An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning (ICML 2023 · similarity 0.88)
- Efficient Exploration and Discriminative World Model Learning with an Object-Centric Abstraction (ICLR 2025 · similarity 0.87)
- Entity-Centric Reinforcement Learning for Object Manipulation from Pixels (ICLR 2024 · similarity 0.85)
- Learning Causal Dynamics Models in Object-Oriented Environments (ICML 2024 · similarity 0.83)
- Tina: Tiny Reasoning Models via LoRA (ICLR 2026 · similarity 0.83)
- Learning Interactive World Model for Object-Centric Reinforcement Learning (NeurIPS 2025 · similarity 0.83)
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.
