Mathematics for AI and Machine Learning

Foundations for modern AI and machine learning

Fisher Information

Chapter 14 Optimization

Chapter 14: Information Theory — Info Geometry

High-resolution PNG
Fisher Information — high-resolution mind-map icon

From the book

Chapter 14: Information Theory. In the chapter mind map this icon labels Info Geometry: Fisher Info Metric. The discussion below is excerpted and lightly edited from § Definition: Fisher Information (Scalar) in Mathematics for AI and Machine Learning.

For a parametric probability distribution $p_{\theta}(x)$ with a single parameter $\theta \in \mathbb{R}$, the Fisher information is a scalar that measures the amount of information that data provides about the parameter:

Intuitively, the Fisher information measures how sensitive the log-probability of the data is to changes in the parameter $\theta$. Higher values indicate that the parameter can be estimated more precisely.

What this drawing shows

What you see. Represents curvature of the log-likelihood or score variance, measuring how much data informs a parameter.

In the mind map. Chapter 14 — Info Geometry: Fisher Info Metric. See From the book above for definitions, figures, and worked examples.

Where to read next

Open Chapter 14 companion →

Read the full definitions, figures, and worked examples in Chapter 14: Information Theory — see the mind-map node Info Geometry: Fisher Info Metric.