We propose a new perspective on deep ReLU networks, namely as circuit counterparts of Lukasiewicz infinite-valued logic -- a many-valued (MV) generalization of Boolean logic. An algorithm for extracting formulae in MV logic from deep ReLU networks is presented. As the algorithm applies to networks with general, in particular also real-valued, weights, it can be used to extract logical formulae from deep ReLU networks trained on data.

翻译：我们提出了深度ReLU网络的新视角，即将其视为卢卡西维茨无限值逻辑（布尔逻辑的多值推广）的电路对应物。本文提出了一种从深度ReLU网络中提取多值逻辑公式的算法。由于该算法适用于具有一般权重（特别包括实值权重）的网络，因此可用于从经数据训练的深度ReLU网络中提取逻辑公式。