AI foundations
Neural network
A layered mathematical model that learns relationships between inputs and outputs.
Definition
A neural network contains interconnected numerical units whose parameters are adjusted during training to transform inputs into useful representations and predictions.
Why it matters
Its behaviour emerges from learned weights rather than explicit business rules, affecting explainability, validation and change control.
Related concepts
- Deep learning
Machine learning based on neural networks with many processing layers.
- Model parameters
The learned numerical values that determine a model's behaviour.
- Model training
The process of adjusting model parameters using examples and an objective.