Research map / Reinforcement learning / Visual and representation RL
Pontryagin Differentiable Programming: An End-to-End Learning and Control Framework
Wanxin Jin, Zhaoran Wang, Zhuoran Yang, Shaoshuai Mou
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 PDP instead of policy cloning or inverse KKT | Imitation learning · Imitation learning environments | training loss | Improved |
| Use PDP instead of policy cloning or inverse KKT | Imitation learning · Imitation learning environments | convergence speed | Improved |
| Learn the control objective with PDP | Robot arm imitation learning · Robot arm | imitation loss | Improved |
| Use PDP instead of a neural-network dynamics model or DMDc | System identification · Cart-pole, robot arm, and quadrotor | training loss | Improved |
| Use PDP instead of a neural-network dynamics model or DMDc | System identification · Cart-pole, robot arm, and quadrotor | convergence speed | Improved |
| Use PDP instead of Adam to train the neural dynamics model | Robot arm neural dynamics learning · Robot arm | training loss | Improved |
| Use PDP instead of Adam to train the neural dynamics model | Robot arm neural dynamics learning · Robot arm | learning speed | Improved |
| Use PDP instead of guided policy search | Feedback control · Cart-pole and robot arm | control loss | Improved |
| Optimize a polynomial policy with PDP instead of using iLQR | Quadrotor planning · Quadrotor | control loss | Mixed |
| Increase polynomial policy degree from N=5 to N=35 (ablation) | PDP polynomial policy · Quadrotor planning | control loss | Improved |
| Use PDP rather than iLQR | Quadrotor planning · Quadrotor | convergence speed | Hurt |
| Use PDP rather than iLQR | Quadrotor planning · Quadrotor | runtime | Improved |
Closest papers by meaning
- Differentiable Information Enhanced Model-Based Reinforcement Learning (AAAI 2025 · similarity 0.84)
- Deterministic Policy Gradient Primal-Dual Methods for Continuous-Space Constrained MDPs (AAAI 2025 · similarity 0.83)
- Constrained Meta-Reinforcement Learning for Adaptable Safety Guarantee with Differentiable Convex Programming (AAAI 2024 · similarity 0.83)
- Free from Bellman Completeness: Trajectory Stitching via Model-based Return-conditioned Supervised Learning (ICLR 2024 · similarity 0.83)
- Goal-Conditioned On-Policy Reinforcement Learning (NeurIPS 2024 · similarity 0.82)
- Neural Time-Reversed Generalized Riccati Equation (AAAI 2024 · similarity 0.82)
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.
