AI RESEARCH
Integrating Lagrangian Neural Networks into the Dyna Framework for Reinforcement Learning
arXiv CS.LG
•
ArXi:2603.08468v1 Announce Type: cross Model-based reinforcement learning (MBRL) is sample-efficient but depends on the accuracy of the learned dynamics, which are often modeled using black-box methods that do not adhere to physical laws. Those methods tend to produce inaccurate predictions when presented with data that differ from the original