Research map / Reinforcement learning / Policy gradient and actor-critic
Guided Policy Optimization under Partial Observability
Yueheng Li, Guangming Xie, Zongqing Lu
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 RL auxiliary loss to learner supervision (GPO-penalty/clip vs GPO-naive) (ablation) | GPO · Brax POMDP tasks | normalized return | Improved |
| constrain the guider to the learner's imitable region (GPO-naive vs PPO+BC) (ablation) | GPO · Brax POMDP tasks | normalized return | Improved |
| collect data with the guider as behavior policy (GPO-ablation vs PPO-asym) (ablation) | PPO with asymmetric value · Brax Humanoid | normalized return | Improved |
| use a larger KL threshold in GPO-penalty (ablation) | GPO-penalty · POPGym CountRecallEasy and BattleshipEasy | return | Improved |
| use a larger KL threshold in GPO-penalty (ablation) | GPO-penalty · POPGym CountRecallHard and BattleshipMedium | return | Hurt |
| add a pre-trained privileged teacher with ADVISOR or PPO+BC | PPO · partially observable Ant | return | No clear effect |
| add asymmetric value function input to PPO | PPO · POPGym memory tasks | return | Mixed |
Closest papers by meaning
- Dual Critic Reinforcement Learning under Partial Observability (NeurIPS 2024 · similarity 0.89)
- Provable Partially Observable Reinforcement Learning with Privileged Information (NeurIPS 2024 · similarity 0.89)
- PIGDreamer: Privileged Information Guided World Models for Safe Partially Observable Reinforcement Learning (ICML 2025 · similarity 0.88)
- Diffusion Policy Policy Optimization (ICLR 2025 · similarity 0.88)
- Provable Representation with Efficient Planning for Partially Observable Reinforcement Learning (ICML 2024 · similarity 0.88)
- Mitigating Partial Observability in Sequential Decision Processes via the Lambda Discrepancy (NeurIPS 2024 · similarity 0.88)
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.
