Research map / Reinforcement learning / Policy gradient and actor-critic
The Power of Learned Locally Linear Models for Nonlinear Policy Optimization
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 |
|---|---|---|---|
| Add synthesized feedback gains to the learned locally linear optimization method (ablation) | Proposed method · Pendulum swing-up and 2D quadrotor stabilization | cost suboptimality | Helped |
| Collect more trajectory data for learned-dynamics iLQR (ablation) | iLQR baselines · Pendulum swing-up and 2D quadrotor stabilization | cost suboptimality | No clear effect |
| Use Opt data collection with Jacobian supervision (ablation) | Learned-dynamics iLQR · Pendulum swing-up and 2D quadrotor stabilization | cost suboptimality | Helped |
| Use Rand rather than Opt data collection with Jacobian supervision (ablation) | Learned-dynamics iLQR with JacReg · Pendulum swing-up and 2D quadrotor stabilization | cost suboptimality | Hurt |
| Replace learned-dynamics iLQR with the proposed method | Pendulum swing-up · Pendulum swing-up | cost suboptimality | Helped |
| Replace learned-dynamics iLQR with the proposed method | 2D quadrotor stabilization · 2D quadrotor stabilization | cost suboptimality | Hurt |
Closest papers by meaning
- Online Policy Gradient for Model Free Learning of Linear Quadratic Regulators with $\sqrt$T Regret (ICML 2021 · similarity 0.89)
- Stabilizing Dynamical Systems via Policy Gradient Methods (NeurIPS 2021 · similarity 0.88)
- Optimal Exploration for Model-Based RL in Nonlinear Systems (NeurIPS 2023 · similarity 0.88)
- Learning convex bounds for linear quadratic control policy synthesis (NeurIPS 2018 · similarity 0.88)
- Truly Deterministic Policy Optimization (NeurIPS 2022 · similarity 0.88)
- DiffOP: Reinforcement Learning of Optimization-Based Control Policies via Implicit Policy Gradients (AAAI 2026 · similarity 0.87)
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.