Research map / Reinforcement learning / Policy gradient and actor-critic
Generalized Off-Policy Actor-Critic
Shangtong Zhang, Wendelin Boehmer, Shimon Whiteson
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 |
|---|---|---|---|
| Optimize the counterfactual objective with Geoff-PAC instead of the excursion objective with ACE | Two-circle MDP · Two-circle MDP | target-policy solution | Improved |
| Increase the counterfactualness parameter hat-gamma (ablation) | Geoff-PAC · Two-circle MDP | final solution | Mixed |
| Vary lambda_2 (ablation) | Geoff-PAC · Two-circle MDP | final solution | No clear effect |
| Use Geoff-PAC instead of ACE or Off-PAC | Off-PAC and ACE · Mujoco robot simulation tasks | J_pi | Mixed |
| Use Geoff-PAC instead of DDPG | DDPG · Hopper and Swimmer | J_pi | Improved |
| Use Geoff-PAC instead of TD3 | TD3 · Hopper and Swimmer | J_pi | Improved |
| Use a uniformly random behavior policy with DDPG | DDPG · HalfCheetah, Walker, and Hopper | J_pi | Hurt |
| Vary lambda_1 (ablation) | ACE · HalfCheetah | performance | No clear effect |
| Use lambda_1 = 0.7, lambda_2 = 0.6, and hat-gamma = 0.2 (ablation) | Geoff-PAC · HalfCheetah | performance | Improved |
Closest papers by meaning
- An Off-policy Policy Gradient Theorem Using Emphatic Weightings (NeurIPS 2018 · similarity 0.92)
- Off-Policy Actor-Critic with Emphatic Weightings (JMLR 2023 · similarity 0.92)
- Offline-Boosted Actor-Critic: Adaptively Blending Optimal Historical Behaviors in Deep Off-Policy RL (ICML 2024 · similarity 0.88)
- Seizing Serendipity: Exploiting the Value of Past Success in Off-Policy Actor-Critic (ICML 2024 · similarity 0.88)
- Dual RL: Unification and New Methods for Reinforcement and Imitation Learning (ICLR 2024 · similarity 0.87)
- Offline Actor-Critic Reinforcement Learning Scales to Large Models (ICML 2024 · similarity 0.87)
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.
