Research map / Reinforcement learning / Policy gradient and actor-critic
Sequential Classification-Based Optimization for Direct Policy Search
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 |
|---|---|---|---|
| update the model after every sample using reused historical solutions (sequential) instead of batch updates | RACOS · Helicopter hovering | top-5 avg hovering steps: 1,477±535 -> 1,837±364 | Helped |
| sequential instead of batch RACOS | RACOS · HalfCheetah | reward: 27961.18±7493.08 -> 36719.90±8288.84 | Helped |
| sequential instead of batch RACOS | RACOS · Ant | reward: 1215.28±1487.81 -> 2114.14±2290.57 | Helped |
| sequential instead of batch RACOS | RACOS · Acrobot | steps (lower better): 169.70±14.15 -> 156.60±18.48 | Helped |
| use random-replace or largest-margin instead of worst-replace for negative set in SRACOS (ablation) | SRACOS · Sphere/Ackley n=100,1000 | function value convergence | No clear effect |
| sequential instead of batch RACOS | RACOS · Sphere/Ackley | function value convergence | Helped |
| sequential instead of batch RACOS | RACOS · Ackley n=100/1000 | wall-clock time per evaluation | Hurt |
Closest papers by meaning
- Trajectory-Based Off-Policy Deep Reinforcement Learning (ICML 2019 · similarity 0.87)
- Efficient Discovery of Pareto Front for Multi-Objective Reinforcement Learning (ICLR 2025 · similarity 0.86)
- Direct Policy Gradients: Direct Optimization of Policies in Discrete Action Spaces (NeurIPS 2020 · similarity 0.86)
- Last-Iterate Global Convergence of Policy Gradients for Constrained Reinforcement Learning (NeurIPS 2024 · similarity 0.86)
- Identifying Policy Gradient Subspaces (ICLR 2024 · similarity 0.85)
- Consensus Based Stochastic Optimal Control (ICML 2025 · similarity 0.85)
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.