Research map / Reinforcement learning / Policy gradient and actor-critic
An Off-policy Policy Gradient Theorem Using Emphatic Weightings
Ehsan Imani, Eric Graves, Martha White
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 emphatic state weighting to the policy gradient | ACE versus OffPAC · Three-state aliased MDP | objective | Improved |
| Use the ordinary behavior-state weighting instead of emphatic weighting (λa = 0) (ablation) | ACE · Three-state aliased MDP | objective | Hurt |
| Set λa to 0.25 rather than using full emphatic weighting (ablation) | ACE · Three-state aliased MDP | objective | Mixed |
| Use λa = 0.5 rather than λa = 0 (ablation) | ACE · Three-state aliased MDP | objective | Improved |
| Use lower λa values (ablation) | ACE · Three-state aliased MDP | learning speed | Mixed |
| Use a GTD(λ) critic rather than exact value estimates (ablation) | ACE · Three-state aliased MDP | objective | Mixed |
| Estimate emphatic weightings online rather than using exact current-policy weightings (ablation) | ACE · 11-state aliased MDP | objective | Hurt |
| Add true emphatic weighting to DPG | DPG · Continuous-action MDP | performance | Improved |
| Use estimated emphatic weightings in ACE rather than the exact weighting (ablation) | ACE · Continuous-action MDP | performance | Hurt |
| Use a stochastic policy rather than a deterministic policy (ablation) | ACE versus True-DPGE · Continuous-action MDP | performance | Hurt |
Closest papers by meaning
- Off-Policy Actor-Critic with Emphatic Weightings (JMLR 2023 · similarity 0.96)
- ACE: Off-Policy Actor-Critic with Causality-Aware Entropy Regularization (ICML 2024 · similarity 0.89)
- Actor-Free Continuous Control via Structurally Maximizable Q-Functions (NeurIPS 2025 · similarity 0.88)
- A Theoretical Justification for Asymmetric Actor-Critic Algorithms (ICML 2025 · similarity 0.88)
- Overestimation, Overfitting, and Plasticity in Actor-Critic: the Bitter Lesson of Reinforcement Learning (ICML 2024 · similarity 0.88)
- Iteratively Refined Behavior Regularization for Offline Reinforcement Learning (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.
