Research map / Reinforcement learning / Policy gradient and actor-critic
Ranking Policy Gradient
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 RPG with off-policy supervised learning instead of the state-of-the-art baselines | RPG · Atari 2600 games (OpenAI Gym) | sample efficiency | Helped |
| Use RPG instead of model-free state-of-the-art methods | RPG · Robotank | final performance | Helped |
| Use RPG instead of model-free state-of-the-art methods | RPG · DoubleDunk | final performance | Helped |
| Use RPG instead of model-free state-of-the-art methods | RPG · Pitfall | final performance | Helped |
| Use RPG instead of model-free state-of-the-art methods | RPG · Pong | final performance | Helped |
| Replace listwise policy learning with pairwise ranking policy learning (ablation) | RPG versus EPG/LPG · Pong | sample efficiency | Helped |
| Use off-policy supervised learning instead of on-policy PPO (ablation) | EPG, LPG, and RPG · Pong | sample efficiency | Helped |
| Use the proposed off-policy supervised learning framework instead of ACER (ablation) | RPG · Pong | sample efficiency | Helped |
| Increase the trajectory reward threshold c (ablation) | RPG · DoubleDunk, BreakOut, BankHeist | sample efficiency | Hurt |
| Increase the trajectory reward threshold c (ablation) | RPG · DoubleDunk, BreakOut, BankHeist | final performance | Helped |
| Use IQN rather than other baselines for exploration | Exploration strategy · Pong | exploration efficiency | Helped |
| Increase exploration efficiency without improving learning efficiency | Rainbow and DQN · Pong | sample efficiency | No clear effect |
Closest papers by meaning
- Statistically Efficient Off-Policy Policy Gradients (ICML 2020 · similarity 0.91)
- Actor-Critics Can Achieve Optimal Sample Efficiency (ICML 2025 · similarity 0.90)
- Generalizable Policy Improvement via Reinforcement Sampling (Student Abstract) (AAAI 2024 · similarity 0.89)
- Kernel Metric Learning for In-Sample Off-Policy Evaluation of Deterministic RL Policies (ICLR 2024 · similarity 0.89)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (NeurIPS 2024 · similarity 0.89)
- Adversarial Policy Optimization for Offline Preference-based Reinforcement Learning (ICLR 2025 · similarity 0.89)
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.