QuestionQ119

VerifiedML – Neural Networks and Testing

A neuron's activation value in a neural network is produced by applying a computation to that neuron.

Which ONE of the following options BEST describes the inputs used to calculate the activation value?

Explanation

A standard neuron computes an activation by applying an activation function to a weighted sum of the preceding layer’s activation values plus its own bias: (a = f(\sum_i w_i x_i + b)). Thus the computation uses the previous-layer activations, the weights on their connections, and a bias for the neuron. TensorFlow documents the equivalent dense-layer operation as activation applied to the input–weight product plus bias.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!