NN2Poly Package

The NN2Poly package focuses on enhancing the interpretability of neural networks by converting them into equivalent polynomial representations. Current research emphasizes methods for extracting rule lists or polynomial approximations from trained networks, addressing challenges in handling various activation functions and network depths, and exploring techniques to improve the efficiency of this conversion process, particularly for large networks. This work contributes to the broader field of Explainable AI (XAI) by providing tools to understand the decision-making process within complex neural networks, thereby increasing trust and facilitating the application of these models in sensitive domains.