liveBook · Manning

Feedback

As explained in Machines that Think

Feedback is information about the effects or outputs of a system that can influence subsequent behavior or processing.

Feedback is information about a system’s effects or outputs that can shape what happens next. In learning systems, it may be sparse, delayed, or noisy rather than complete and immediate. section 5.3.5

Treating limited feedback as a basic condition has encouraged local learning rules, continual adjustment, and credit assignment that does not depend on globally propagated error signals. Brain modelling highlights these constraints, but does not provide a blueprint for building thinking machines. section 5.3.5

In human–machine interaction, feedback can contribute to dynamics that resemble comprehension, but fluent or responsive behavior alone does not settle whether a system understands. In education, AI assistants can offer feedback, dialogue, and practice while also making learner-produced work less reliable as evidence of understanding. section 1.6.2 section 8.3

How Machines that Think covers it

How deep it goes
Explains sparse, delayed, and noisy feedback and names resulting learning approaches, but stops short of implementation detail.
Examples
Contrasts sparse, delayed, noisy biological feedback with AI assistants’ educational feedback, dialogue, and practice.
What you can do after
Evaluate how feedback can shape learning and interaction without treating responsive output as proof of understanding.
How it connects
Connects feedback to learning, credit assignment, continual learning, adaptive control, and human–machine interaction.

Also explained in

Compare how 25 books explain feedback →

Related concepts

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