Mathematics for AI and Machine Learning

Foundations for modern AI and machine learning

Bellman Backup

Chapter 15 Reinforcement learning

Chapter 15: Bellman Equations and Operators — Policies

Animated preview (GIF)
Bellman Backup — animated GIF preview
High-resolution PNG
Bellman Backup — high-resolution mind-map icon

From the book

Chapter 15: Bellman Equations and Operators. In the chapter mind map this icon labels Policies: Stochastic $\pi(a|s)$ & Value $V(s)$. The discussion below is excerpted and lightly edited from § Definition: State Value Function in Mathematics for AI and Machine Learning.

The state value function $V_\pi: \mathcal{S} \to \mathbb{R}$ under policy $\pi$ is the expected return starting from state $s$:

where the expectation is over trajectories generated by following policy $\pi$.

What this drawing shows

What you see. Root state $s$ fixed; successor backups from $s_1$ and $s_2$ grow upward, illustrating the Bellman backup tree.

In the mind map. Chapter 15 — Policies: Stochastic & Value. See From the book above for definitions, figures, and worked examples.

Where to read next

Open Chapter 15 companion →

Read the full definitions, figures, and worked examples in Chapter 15: Bellman Equations and Operators — see the mind-map node Policies: Stochastic & Value.