What is Weight initialization? Definition and meaning
Weight initialization is the choice of starting values for a neural network's weights and biases before training begins.
Definition
Weight initialization: Weight initialization is the choice of starting values for a neural network's weights and biases before training begins.
In short: A good start matters.
When it is used
Every neural network needs it, and the right scheme makes deep networks trainable and stable.
Where it fits
Weight initialization belongs to Neural network basics in the Deep learning topic. These are the other terms in the same group:
- Neural networks: A neural network is a model made of layers of simple units that each compute a weighted sum and a non-linear function, with the weights learned from data.
- Activation functions: An activation function is the non-linear function applied to a neuron's weighted sum before its output is passed on.
- Learning rate schedules: A learning rate schedule is a rule that changes the optimiser's learning rate as training progresses.
Learn more
The full guide to Weight initialization explains how it works, when to use it and how it compares with the alternatives, with examples. The glossary lists every term on the site from A to Z.
Previous term: Weaviate · Next term: Weights & Biases