liveBook · Manning

Bias

As explained in AI Model Evaluation

A bias is a learnable additive parameter that shifts a model computation independently of its inputs.

A bias is a learnable additive parameter that shifts a model computation independently of its inputs. Alongside weights, it helps a model adjust as it learns from data. section 1.3

This parameter is distinct from bias in predictions: systematic errors or disparities that can disadvantage particular groups or scenarios. Diagnostic evaluations can help identify those disparities. section 2.1.2

How AI Model Evaluation covers it

How deep it goes
Mentions biases alongside weights but does not explain their calculation or training mechanics.
Examples
Names personalized recommendations as a model application without illustrating a bias parameter.
What it assumes
Assumes familiarity with basic model anatomy and points readers elsewhere for model-design details.
How it connects
Connects biases with weights, feature inputs, activation functions, and neural network layers.

Also explained in

Compare how 38 books explain bias →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Model Evaluation. The section links point to the parts of the book it draws on.