Research map / Reinforcement learning / Policy gradient and actor-critic
Sample Efficient Policy Gradient Methods with Recursive Variance Reduction
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 recursive variance-reduced policy gradients with step-wise importance weights instead of SVRPG | Policy gradient optimization · Nonconcave policy optimization; epsilon-approximate stationary point | trajectory complexity: O(1/ε^{5/3}) -> O(1/ε^{3/2}) | Helped |
| Use recursive variance-reduced policy gradients instead of standard stochastic policy gradients | Policy gradient optimization · Nonconcave policy optimization; epsilon-approximate stationary point | trajectory complexity: O(1/ε^2) -> O(1/ε^{3/2}) | Helped |
| Choose inner-loop batch size B around the order of sqrt(N) (ablation) | SRVR-PG · Pendulum | convergence performance | Helped |
| Increase inner-loop batch size beyond the best-performing setting (ablation) | SRVR-PG · Pendulum | sample efficiency | Hurt |
| Extend recursive variance reduction to parameter-based exploration | PGPE · Classic reinforcement-learning benchmarks | performance | Helped |
Closest papers by meaning
- On the Convergence and Sample Efficiency of Variance-Reduced Policy Gradient Method (NeurIPS 2021 · similarity 0.93)
- Policy Optimization with Stochastic Mirror Descent (AAAI 2022 · similarity 0.91)
- PAGE-PG: A Simple and Loopless Variance-Reduced Policy Gradient Method with Probabilistic Gradient Estimation (ICML 2022 · similarity 0.90)
- Reusing Trajectories in Policy Gradients Enables Fast Convergence (ICML 2026 · similarity 0.90)
- Learning Optimal Deterministic Policies with Stochastic Policy Gradients (ICML 2024 · similarity 0.90)
- Convergence Analysis of Policy Gradient Methods with Dynamic Stochasticity (ICML 2025 · similarity 0.90)
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.