ORLA is a symbolic reinforcement learning approach that learns a value-based argumentation framework as a reasoning engine for solving a task. This repo demonstrates ORLA on both the Foggy Frozen Lake and the Takeaway tasks.
reinforcement-learning deep-reinforcement-learning argumentation explainable-ai symbolic-reinforcement-learning frozen-lake keepaway orla
-
Updated
Aug 23, 2023 - Jupyter Notebook