Research map / Reinforcement learning / Policy gradient and actor-critic
Expected Policy Gradients
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 |
|---|---|---|---|
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · HalfCheetah-v1 | learning performance | Helped |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · InvertedPendulum-v1 | learning performance | Helped |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · Reacher2d-v1 | learning performance | Helped |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · Walker2d-v1 | learning performance | Helped |
| Use stochastic policy gradients with OU exploration and constant diagonal covariance | SPG · HalfCheetah-v1 | learning performance | Hurt |
| Use stochastic policy gradients with OU exploration and constant diagonal covariance | SPG · Reacher2d-v1 and Walker2d-v1 | learning performance | Hurt |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · InvertedPendulum-v1 | estimated standard deviation across runs after learning: 291.26 [241.45, 351.88] -> 0.00 n/a | Helped |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · Reacher2d-v1 | estimated standard deviation across runs after learning: 1.22 [0.63, 2.31] -> 0.13 [0.07, 0.26] | Helped |
| Replace OU exploration with Hessian-exponential Gaussian exploration | DPG, yielding EPG · Walker2d-v1 | estimated standard deviation across runs after learning: 543.54 [450.58, 656.65] -> 762.35 [631.98, 921.00] | Hurt |
Closest papers by meaning
- Fourier Policy Gradients (ICML 2018 · similarity 0.92)
- Learning Optimal Deterministic Policies with Stochastic Policy Gradients (ICML 2024 · similarity 0.91)
- Convergence Analysis of Policy Gradient Methods with Dynamic Stochasticity (ICML 2025 · similarity 0.90)
- PC-PG: Policy Cover Directed Exploration for Provable Policy Gradient Learning (NeurIPS 2020 · similarity 0.90)
- Last-Iterate Global Convergence of Policy Gradients for Constrained Reinforcement Learning (NeurIPS 2024 · similarity 0.88)
- Global Optimality for Constrained Exploration via Penalty Regularization (NeurIPS 2026 · 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.