Research map / Reinforcement learning / Policy gradient and actor-critic
Learning Value Functions in Deep Policy Gradients using Residual Variance
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 |
|---|---|---|---|
| Train the critic with residual variance instead of conventional MSE | SAC · MuJoCo and PyBullet continuous-control tasks | average total reward: +26% on average | Helped |
| Train the critic with residual variance instead of conventional MSE | PPO · MuJoCo and PyBullet continuous-control tasks | average total reward: +39% on average | Helped |
| Train the critic with residual variance instead of conventional MSE | TRPO · MuJoCo and PyBullet continuous-control tasks | average total reward | Mixed |
| Train the critic with residual variance instead of conventional MSE | PPO · AntBullet and HalfCheetahBullet | MSE to empirical value target | Hurt |
| Train the critic with residual variance instead of conventional MSE | PPO · AntBullet and HalfCheetahBullet | error to true value function | Helped |
| Train the critic with residual variance instead of conventional MSE | PPO · continuous-control experiments | gradient variance | Helped |
| Train the critic with residual variance instead of conventional MSE | PPO and TRPO · Acrobot and MountainCar | average total reward | Helped |
| Vary the bias and residual-variance weighting using L_alpha or L_(1/alpha) (ablation) | AVEC-PPO · PyBullet continuous-control tasks | average total reward | Mixed |
Closest papers by meaning
- Value Improved Actor Critic Algorithms (NeurIPS 2025 · similarity 0.91)
- ReLU to the Rescue: Improve Your On-Policy Actor-Critic with Positive Advantages (ICML 2024 · similarity 0.90)
- Actor-Free Continuous Control via Structurally Maximizable Q-Functions (NeurIPS 2025 · similarity 0.89)
- Mean Field Langevin Actor-Critic: Faster Convergence and Global Optimality beyond Lazy Learning (ICML 2024 · similarity 0.89)
- Wasserstein Policy Optimization (ICML 2025 · similarity 0.88)
- Efficient Off-Policy Learning for High-Dimensional Action Spaces (ICLR 2025 · 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.