Artificial Neural Network
Definition
An artificial neural network is a parameterizable computational model composed of connected layers or units.
Weights are typically learned by optimization from data. Architecture, activation functions, and training methods determine which functions the network can approximate.
Simply explained
Like many small computational building blocks whose connections are jointly adapted from examples.
Why it matters in practice
Basis of many modern image, speech, audio, and forecasting models.
Example from business
A network learns from sensor data whether a production state is normal or anomalous.
Limits and pitfalls
- Overstretching biological analogy; ignoring training data and generalization; equating network size with quality.
Origin & history
Neural networks have roots in works by McCulloch/Pitts, Rosenblatt, and backpropagation research; today’s deep learning networks build on these.
Use in mid-sized companies
- Possible application in SMEs: Choose as a model class if data volume, task, and operating costs justify the effort.
Frequently asked questions
Is a neural network a brain model?
Only very loosely inspired. Is deep learning the same? Deep learning typically uses deep neural networks.
Sources
- Deep Learning · Goodfellow et al., Deep Learning
Sources checked on 10.10.2026.
Editorially reviewed on: