hAIthem Framework for Physics Discovery via RL and LLMs

The hAIthem framework combines reinforcement learning with LLM-derived literature knowledge to search high-dimensional Lagrangian parameter spaces. It uses an RL agent to play a Battleship-style game to identify theory signatures not excluded by current experimental constraints.