Skip to content
    Flow-alerta.cab

    Activation Function

    AI fundamentals & modelsNeural activation function
    Editorially reviewed on: 1 source
    Your reading view

    The short answer · 9 words

    In short

    Mathematical function for converting signals in a neural network.

    Flow-alerta.cab (2026): Activation Function. https://flow-alerta.cab/en/glossar/activation-function. Editorially reviewed on: 09 October 2026.

    How much detail would you like?

    Practical

    Helps you understand the structure and training behavior of neural networks.

    Example scenario: A hidden layer uses ReLU to set negative input values to zero.

    Choose consciously for own model projects; for finished AI services usually already defined by the model.

    At a glance

    Core
    An activation function transforms the calculated signal of a neuron, for example its weighted input sum. It is usually non-linear so that several layers can depict complex relationships. Examples are ReLU, Sigmoid and GELU; the appropriate function depends on architecture and task.
    Use
    Helps you understand the structure and training behavior of neural networks.
    Example
    Example scenario: A hidden layer uses ReLU to set negative input values to zero.
    Limits
    Interpret each activation as probability; choose the same output function for all tasks; overlook unfavorable gradients.

    In context

    Backpropagation

    Efficient method for calculating gradients in multilayer neural networks.

    Questions about this term

    Are all activations non-linear?

    No, there is also the linear identity function.

    Is each output value between zero and one?

    No, it depends on the function.

    All details and background

    An activation function transforms the calculated signal of a neuron, for example its weighted input sum. It is usually non-linear so that several layers can depict complex relationships. Examples are ReLU, Sigmoid and GELU; the appropriate function depends on architecture and task.

    ELI5 – Explained simply

    A computation rule determines how an intermediate result is passed on to the next stage.

    Why it matters in practice

    Helps you understand the structure and training behavior of neural networks.

    Example

    Example scenario: A hidden layer uses ReLU to set negative input values to zero.

    Pitfalls and limits

    • Interpret each activation as probability; choose the same output function for all tasks; overlook unfavorable gradients.

    Origin & history

    Activation functions are among the basic building blocks of neural networks. Keras documents both classical functions such as sigmoid and variants such as GELU.

    Use in mid-sized companies

    • Choose consciously for own model projects; for finished AI services usually already defined by the model.

    Related terms

    Matching Flow-alerta service

    AI opportunity analysis & roadmapAI consulting: AI potential analysis & roadmap on tasks, data and prerequisites of a possible first AI use case.

    Matching tool / use case

    View use casesTool example: Activation functions in Keras.

    Sources

    • keras.ioKeras, Layer activation functions

    Quellenprüfung: 09.10.2026. Die Quellen belegen Definition und dokumentierte Einordnung; Anwendungsbeispiele sind mögliche Szenarien.

    Editorially reviewed on:

    Next step

    Put knowledge into practice

    Discuss your projectView use cases