News for hackers. Stuff that matters.
Re:Neural Networks Are Decision Trees by groar on Monday October 17, 2022 @07:18AM (#33232043)
The author shows that any neural network having piece-wise linear activation functions can be represented as a decision tree
Re:Neural Networks Are Decision Trees by Rochus on Monday October 17, 2022 @08:19AM (#33232521)
That's interesting. Is it still that the Rectified Linear Unit (ReLU) is the prevailing activation function in deep neural networks, because of the the vanishing gradients with activation functions like tanh? If so the conclusions from the paper would apply to a very wide range of deep neural networks.