Neural Networks are Decision Trees
This work addresses the black-box nature of neural networks by providing a foundational equivalence for interpretability, though it is incremental in scope.
The authors demonstrated that any neural network can be exactly represented as a decision tree, preserving accuracy, which offers interpretability and computational benefits for small networks.
In this manuscript, we show that any neural network with any activation function can be represented as a decision tree. The representation is equivalence and not an approximation, thus keeping the accuracy of the neural network exactly as is. We believe that this work provides better understanding of neural networks and paves the way to tackle their black-box nature. We share equivalent trees of some neural networks and show that besides providing interpretability, tree representation can also achieve some computational advantages for small networks. The analysis holds both for fully connected and convolutional networks, which may or may not also include skip connections and/or normalizations.