Research map / Learning theory and algorithms / Combinatorial and search optimization
Safe Pontryagin Differentiable Programming
Wanxin Jin, Shaoshuai Mou, George J. Pappas
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 barrier-based Strategy (B) to approximate both the trajectory and its gradient | Safe PDP for learning MPCs from demonstrations · Demonstration reproducing loss across the four systems | reproducing loss | Improved |
| Use the constrained solver for the trajectory and Safe PDP differentiation for its gradient (Strategy C) (ablation) | Learning MPCs from demonstrations · Demonstration reproducing loss across the four systems | reproducing loss | Improved |
| Use direct constrained differentiation with active-constraint identification (Strategy A) (ablation) | Learning MPCs from demonstrations · Demonstration reproducing loss across the four systems | optimization stability | Hurt |
| Set the outer barrier parameter to epsilon <= 10^{{-}2} | Safe neural policy optimization · Robot arm and 6-DoF maneuvering quadrotor | control-cost convergence | Improved |
| Set the outer barrier parameter to epsilon <= 10^{{-}2} | Safe motion planning · Cartpole and 6-DoF rocket landing | task-loss convergence | Improved |
| Use barrier functions to approximate constrained trajectories and gradients | Safe PDP · Safe policy optimization and safe motion planning | constraint satisfaction during optimization | Improved |
Closest papers by meaning
- A Physics-Informed Machine Learning Framework for Safe and Optimal Control of Autonomous Systems (ICML 2025 · similarity 0.85)
- Constrained Meta-Reinforcement Learning for Adaptable Safety Guarantee with Differentiable Convex Programming (AAAI 2024 · similarity 0.85)
- Learning Safe Control via On-the-Fly Bandit Exploration (ICML 2025 · similarity 0.84)
- Feasible Reachable Policy Iteration (ICML 2024 · similarity 0.84)
- Boundary-to-Region Supervision for Offline Safe Reinforcement Learning (NeurIPS 2025 · similarity 0.84)
- Provably Efficient RL for Linear MDPs under Instantaneous Safety Constraints in Non-Convex Feature Spaces (ICML 2025 · similarity 0.83)
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.
