Research map / Detection and segmentation / Open-set and few-shot detection
Unsupervised Video Object Segmentation for Deep Reinforcement Learning
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 |
|---|---|---|---|
| Pretrain the motion pathway with unsupervised object segmentation, then jointly optimize it with the RL objective | A2C · Atari | sample complexity: 26 games improved; 3 games worse | Helped |
| Pretrain the motion pathway with unsupervised object segmentation, then jointly optimize it with the RL objective | PPO · Atari | performance: 25 games benefit; 9 games deteriorate | Mixed |
| Use a randomly initialized MOREL architecture instead of the standard A2C architecture (ablation) | A2C · Atari ablation study | sample complexity | No clear effect |
| Initialize the motion pathway from the object-segmentation network without joint training (ablation) | A2C · Atari ablation study | sample complexity | Helped |
| Initialize from object segmentation and add joint optimization (ablation) | A2C · Q*bert | performance | Helped |
| Continue joint optimization of the segmentation pathway (ablation) | A2C · Beam Rider | performance | No clear effect |
Closest papers by meaning
- Data Augmentation for Instruction Following Policies via Trajectory Segmentation (AAAI 2025 · similarity 0.88)
- Reinforcement Cutting-Agent Learning for Video Object Segmentation (CVPR 2018 · similarity 0.87)
- An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learning (ICML 2023 · similarity 0.87)
- Efficient Exploration and Discriminative World Model Learning with an Object-Centric Abstraction (ICLR 2025 · similarity 0.86)
- Towards Principled Representation Learning from Videos for Reinforcement Learning (ICLR 2024 · similarity 0.86)
- Investigating Pre-Training Objectives for Generalization in Vision-Based Reinforcement Learning (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.