liveBook · Manning

Action-dependent feature

As explained in Optimization Algorithms: AI techniques for design, planning, and control problems

A feature representation whose values are associated with particular actions in a contextual decision problem.

An action-dependent feature (ADF) associates features with individual actions, helping a contextual bandit choose among alternatives using both shared context and action-specific information. section 12.5

Delivery truck selection

For a city context, the model can select a medium electric truck with snow tires, then receive a noisy reward based on the component values.

How Optimization Algorithms: AI techniques for design, planning, and control problems covers it

How deep it goes
Implements action enumeration, PMF sampling, reward feedback, training, and testing without developing formal theory.
Examples
Uses delivery contexts, truck sizes, engine types, tire types, generated action combinations, and simulated noisy rewards.
What you can do after
Apply it via Vowpal Wabbit to generate, explore, train, and test action-specific contextual-bandit decisions.
How it connects
Connects contextual bandits with exploration, reward modeling, probability mass functions, and feature interactions.
Cover of Optimization Algorithms: AI techniques for design, planning, and control problems
liveBookmanning.com
Read Optimization Algorithms: AI techniques for design, planning, and control problems on liveBook Get the book at manning.com

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Optimization Algorithms: AI techniques for design, planning, and control problems. The section links point to the parts of the book it draws on.