liveBook · Manning

Causation

As explained in Inside AI

Causality is the relationship in which one event, condition, or variable produces or influences another.

A prediction can be useful without explaining why something happens. Causation concerns how an event, condition, object, or process influences another. It asks more than whether two events occur together—or which one came first. section 11.22

The heart and blood circulation

The heart’s pumping action causes blood to circulate. A system that finds patterns in medical data may still fail to infer this basic relationship. That limits its ability to trace chains of causes that could help with diagnosis or treatment. section 8.0

Understanding causes is an aim for next-generation AI: a system that can reason about underlying factors could plan and act more effectively than one that only recognizes correlations. Encoding cause and effect in a computer program remains a challenge, however. A predictive result should not be mistaken for a causal explanation. section 11.22 section 8.0

How Inside AI covers it

How deep it goes
Defines cause and effect conceptually and distinguishes it from chronology and correlation, without providing an inference procedure.
Examples
Uses heart pumping and blood circulation to illustrate a causal fact that pattern-based medical AI may miss.
What you can do after
Evaluate/understand why a useful prediction does not necessarily explain a cause.
How it connects
Links causation to correlation, predictive models, medical diagnosis, and AI planning.

Also explained in

Compare how 8 books explain causation →

Related concepts

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