Uses self-entropy/self-attention to learn the first layer in a two layer neural network with a single pass. Self-entropy is calculated as: entropy(softmax(softmax(X*X^T)*X))
Unsupervised feed forward layer learning with a single pass.
Uses self-entropy/self-attention to learn the first layer in a two layer neural network with a single pass. Self-entropy is calculated as: entropy(softmax(softmax(X*X^T)*X))
Unsupervised feed forward layer learning with a single pass.
BSD 3-Clause "New" or "Revised" License