Research map / Reinforcement learning / Policy gradient and actor-critic

Physics-Informed Optimal Control by Control-Only Supervision with Error Guarantees on Value and Policy

NeurIPS 2026 · Reinforcement learning › Policy gradient and actor-critic · approach: Model-based RL with guarantees

Read the paperCheck your idea against 108,967 papers

AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Proposes a physics-informed neural method for high-dimensional optimal control that trains a value network using only control labels plus an HJB residual, with error guarantees on value and policy.

Problem

Solving high-dimensional optimal control problems via HJB equations typically requires value function labels, which are expensive to obtain; the paper asks whether control-only supervision suffices.

Approach

A neural network parameterizes the value function and the policy is derived from its gradient; training minimizes a weighted sum of the HJB residual and the mismatch between the network's control output and the given control labels, so no value labels are needed.

Experimental setup

Builds on: physics-informed neural networks

Closest papers by meaning

• Physics-Informed Neural Network Policy Iteration: Algorithms, Convergence, and Verification (ICML 2024 · similarity 0.90)
• Neural Time-Reversed Generalized Riccati Equation (AAAI 2024 · similarity 0.88)
• Physics-Informed Representation and Learning: Control and Risk Quantification (AAAI 2024 · similarity 0.87)
• Optimal Control Operator Perspective and a Neural Adaptive Spectral Method (AAAI 2025 · similarity 0.86)
• A Physics-Informed Machine Learning Framework for Safe and Optimal Control of Autonomous Systems (ICML 2025 · similarity 0.85)
• Deep learning for continuous-time stochastic control with jumps (NeurIPS 2025 · similarity 0.84)