Research map / Reinforcement learning / Policy gradient and actor-critic
PODS: Policy Optimization via Differentiable Simulation
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 |
|---|---|---|---|
| Analytic-gradient policy improvement (PODS, 2nd order) instead of SAC/PPO/GPS | PODS · Multiple payload manipulation tasks (Table 1) | reward | Helped |
| Use PODS vs BPTT (1st order PODS and 2nd order PODS vs BPTT) (ablation) | PODS 2nd order update · 2D Simple Pendulum, 50-step rollouts | convergence and final reward | Helped |
| Use 2nd order update rule instead of 1st order (ablation) | PODS · 2D Simple Pendulum, 50-step rollouts | convergence rate | Helped |
| Smaller policy-update batch size (500 vs 4000 rollouts) (ablation) | PODS · Cloth laying task | sample efficiency and wall-clock time | Helped |
| Use smaller batch size (500) instead of 4000 for policy update | PODS · Table 1 payload tasks | reward | Mixed |
Closest papers by meaning
- Diffusion Policy Policy Optimization (ICLR 2025 · similarity 0.90)
- Learning Optimal Deterministic Policies with Stochastic Policy Gradients (ICML 2024 · similarity 0.89)
- Wasserstein Policy Optimization (ICML 2025 · similarity 0.89)
- Adaptive-Gradient Policy Optimization: Enhancing Policy Learning in Non-Smooth Differentiable Simulations (ICML 2024 · similarity 0.88)
- Differentiable Information Enhanced Model-Based Reinforcement Learning (AAAI 2025 · similarity 0.88)
- Oracle-Efficient Reinforcement Learning for Max Value Ensembles (NeurIPS 2024 · 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.